A Study of the Chromosomes of the Germ Cells

of Metazoa.

BY THOMAS H.

. JR., PH.D.

READ BEFORE i \MERICAN PHILOSOPHICAL IT, JANUARY 18, 1901.

' *

.4 T <-

A STUDY OF THE CHKOMOSOMES OF THE GERM CELLS OF METAZOA.

Plates IV VIII.
BY THOS. H. MONTGOMERY, JR., PH.D.,

ASSISTANT PROFESSOR OF ZOOLOGY, UNIVERSITY OF PENNSYLVANIA, PHILADELPHIA.

Read January 18, 19O1.

I. INTRODUCTION.

The present study is practically a continuation of previous work of mine upon sper-
matogenesis in the Arthropods. It was undertaken primarily to correct certain errors of
interpretation and observation in my work on Pentatoma (JS/uchistua). But many mor-
phological problems arose in connection with this reexamination, sucli as the significance
of the changes in the synapsis stage, the significance of the chromatin nucleoli, the
reasons for a reduction division, the significance of the sequence of the stages of the
germinal cycle, and the question as to why different species have different numbers of
chromosomes. Thus my investigations given here are essentially on the history of the
chromosomes during the germinal cycle.

It is impossible to answer these problems by an examination of a single species, and
accordingly there are presented here the results of a comparative study of the spermato-
genesis of some forty-two species of Hemiptera heteroptera, belonging to twelve different,
families. This comparative study, has brought to light certain wholly unexpected phe-
nomena, and none less anticipated than the discovery of four species with an uneven
normal number of chromosomes ; this discovery has furnished facts for explaining how
the chromosomal numbers may change with the evolution of the species, and how the
chromatin nucleoli may have originated. And only such a comparative study could
furnish facts to show that in the synapsis stage bivalent chromosomes are formed by the
union of paternal with maternal chromosomes i. e., that this is the stage of conjugation
of the chromosomes. The comparative method in Cytology cannot be overestimated,
though of course careful detailed examinations of single objects should be carried on at
the same time. For a single object is rarely capable of serving as the basis of explana-

.a

A STUDY OF THE CHROMOSOMES OF THE GERM CELLS OF METAZOA. 155

tion of all the problems ; an investigation of a number of forms always shows that some
are more favorable than others for answering certain questions, and then there is the
chance that a wholly unexpected discovery may be made that may have great signifi-
cance. So the plea is made here for the comparative method in Cytology, and Cytology
should not be ranked as a line of work separate from others it is all Morphology in the
broad sense of the term, and it only happens that in Cytology we use higher magnifica-
tion powers of the microscope than in other lines. If one form shows phenomena that
seem inexplicable after careful work, then the proper method, the one that would promise
a surer reaching of results, is not to reexamine this form again and again, but to compare
other forms in the search for the explanation.

In the present paper the part containing the general conclusions may appear dispro-
portionally great to the record of the observations. These observations are to great
extent on the number and valence of the chromosomes and chromatin nucleoli from the
time of the last generation of the spermatogonia up to the formation of the spermatids.
But the determination of these numbers is very difficult ; large numbers of sections have
to be examined in order to find the necessary stages, and the number of the chromosomes
of each stage have to be counted in a considerable number of cells of each species in order
to insure accuracy. The counting has been done in all cases by selecting those cells in
which the chromosomes are most loosely grouped, being sure at the same time that all
the chromosomes are in the plane of the section, drawing the chromosomes carefully with
the camera lucida, then counting their number on the drawings. This demands much
patience and time, necessitating also constant reexaminations and study of new material,
though the results may be tabulated in a very small space. Of course the difficulties are
most pronounced where the chromosomes are numerous and small.

The material was collected by me at two localities in the vicinity of Philadelphia,
Pennsylvania, and in the neighborhood of Wood's Holl, Massachusetts. Great care was
taken to insure accurate identification of the species, and my specimens were directly
compared by me with the collections in the museums of the Wagner Institute of Science
and of the Academy of Natural Sciences of this city ; these collections had been labeled
by Dr. P. R. Uhler, of Baltimore, our foremost American authority on this group of
Insects ; and I must also acknowledge my indebtedness to Dr. Uhler for kindly identify-
ing a number of species which were not represented in the collections just mentioned.
To my friend, Mr. C. W. Johnson, curator of the Wagner Institute, my thanks are also
due for aid in identification. The differences of the spermatogenetic phenomena of
different species shows how important it is to secure accurate identification.

The testes were removed as rapidly as possible from the living animals and immedi-
ately placed in the fixing fluids, Flemming's chronio-aceto-osmic acid mixture (the

1*56 MONTGOMERY A STUDY OF THE CHROMOSOMES

stronger solution), Hermann's chromo-aceto-platinic chloride mixture, and a picro-acetic
mixture recommended by Prof. Conklin (100 parts saturated aqueous solution of picric
acid, 100 parts distilled water, G parts glacial acetic acid) being used. Of these the
mixtures of Flemming and Hermann proved the best for the chromosomal structures, for
the picro-aSetic mixture, while giving an excellent preservation of the actromatic spindle
structures, swells the chromosomes very considerably so that on pole views of monaster
stages they generally appear closely apposed to one another, which makes it difficult to
count them. Where the species is small it is necessary to remove the testes in the fixa-
tive under a dissecting microscope. The sections were stained either by the iron-hsema-
toxylin method of Heidenhain or by the saffranine-gentian violet method of Hermann.

II. OBSERVATIONS.

PENTATOMID^E.

1. Euchistm variolarius Pal. Beauv.

This is the species the spermatogenesis of which I described under the name of
" Pentatoma " in a former paper (1898) ; twenty-eight testes were studied from adult
individuals of all seasons except the winter months.

In my former paper (I. c.) I did not find chromatin nucleoli in the spermatogonia ; I
concluded that there was no stage of longitudinal splitting of the chromosomes during the
growth period, and I concluded that the second maturation division was a reduction
mitosis like the first. Shortly afterward appeared the papers by Paulmier (1898, 18!M>)
on the spermatogenesis of Anasa tristis, wherein he showed that there are two chromatin
nucleoli (his "small chromosomes") in the spermatogonia, and that these unite in the
spermatocytes to form one bivalent one ; that the chromosomes undergo a longitudinal
splitting in the growth period, and that the second maturation division is equational. In
those points wherein I differed from Paulmier, I find that Paulmier is correct, and that I
gave a wrong interpretation to the phenomena in Euchistus. I find nothing to correct in
the matter of the other points described in my earlier account, and here give briefly
merely the necessary emendations to my former paper.

Spermatoyonia. In the resting spermatogonium there are in the nucleus beside the
true nucleolus (of which there may be more than one) two small chromatin nucleoli of
rounded form (Plate I, Fig. 1, N. 2). With the saffranine-gentian violet stein of Her-
mann, when properly used, these stain bright red, the true nucleolus a faint bluish, the
chromatin proper a deep violet ; careful staining and thin sections are necessary to show
them plainly. Sometimes one or both of them are attached to a true nucleolus. In the
prophases of mitosis the chromatin nucleoli are easily recognizable by being muct^ smaller

OF THE GERM CELLS OF METAZOA. 157

and more spherical than the chromosomes. In the monaster stage, in favorable cases
where the chromatic elements are not too densely arranged, are seen fourteen larger
elements, the chromosomes proper, and two smaller ones regularly rounded in form,
which are the chromatin nucleoli (Figs. 2, 3, N. %}. Sometimes the chromosomes are
rounded, but since they frequently appear slightly elongate on pole view of the spindle,
their division in metakinesis must be an equational one. In the metakinesis all sixteen
elements, the fourteen chromosomes and the two chromatin nucleoli, are divided, so that
each daughter cell (first spermatocyte) receives sixteen elements.

Thus there are two chromatin nucleoli in the spermatogoiiia, and the chromatin
nucleolus of the spermatocystes is not, as I had previously described, formed by a modifi-
cation of one of the fourteen chromosomes of the spermatocytes, but is derived from the
two of the spermatogoiiia. My error was perhaps excusable, since in restudying the
preparations which were used for my former paper I find that they are not suitably
stained to show the chromatin nucleoli in the spermatogoiiia.

Growth period of the spermatocytes (anaphases of the last spermatogonic division,
synapsis, postsynapsis, telophase and rest). The fourteen chromosomes in each daughter
cell (first spermatocyte) pass toward the pole of the spindle and become irregular in
contour and form. Then each becomes longitudinally split (Figs. 411). This splitting
cannot be clearly seen in all preparations, and is by no means as clear as in Anasa and
certain other Hemiptera '.; the preparations of my former paper were too deeply stained to
show it. The split commences in the-early synapsis stage (Fig. 4) and is most marked in
the postsynapsis (Fig. 9), and is clearly a single longitudinal split. Never do the split
halves separate widely from .one another, as Paulmier found for Anasa, but always appeal-
to remain close together and approximately parallel ; at the most there is a divergence
only at the ends of the chromosomes. On deep staining the split may be easily over-
looked. The two chromatin nucleoli do not become loose in texture, retain their charac-
teristic red stain with saffranine, and join together in the early synapsis to form one
dumbbell-shaped (bivalent) one (N. 2, Figs. 4, 5, 8, 10) ; they do not become longitudi-
nally split like the chromosomes proper. In the early synapsis they are frequently very
irregular in form, as I showed in my previous paper, but the apparent fragmentation of
them which I then described a fragmentation of a single long one into two is not a
fragmentation at all, but a stage before the two have joined to form one bivalent one.

Reduction in number of the chromosomes. In my earlier paper I showed that the
number of chromosomes is reduced one-half during the syiiapsis period i. e., long before
the maturation divisions. I then considered it probable that the reduction in number
was effected by a union of chromosomes end to end, but was unable to prove this point.
Since then I have been able to demonstrate that this numerical reduction is effected in

l'">8 MONTGOMERY A STUDY OF TJUO CHROMOSOMES

the synapsis by the union into seven pairs of the fourteen chromosomes, each of the
seven bivalent chromosomes (pairs) being composed of two univalent chromosomes joined
end to end (Figs. 5-11). Where the ends of two univalent chromosomes come together
is seen a connecting band of linin ; each bivalent chromosome during the synapsis and
postsynapsis is U- or V-shaped, and the bend or angle of the U or V marks the point of
union of two univalent dhromosomes ; the arms of the U or V are longitudinally split.
In each bivalent chromosome only one end of each univalent chromosome is thus closely
connected with one end of the other, the opposite ends of the univalent chromosomes
having no such linin connections. It has been already mentioned that the two chromatin
nucleoli come together likewise to form one bivalent one, and it can be seen that they are
connected by a band of linin.

In a paper on the spermatogenesis of Peripatm (1901) I showed that it is a particular
end of one univalent chromosome which .unites with a particular end of another; these
ends are the ones which point nearest to the pole of the spindle in the anaphase of the
last sperniatogonic mitosis, the " central ends," as I have called them, in distinction to the
opposite or " distal ends." In Euchistus, on the contrary, I am unable to determine
positively whether it is similar ends of chromosomes which unite, because in this form the
chromosomes have a much more irregular position within the nucleus ; the polarity of the
nucleus is not so well marked as in Peripatus. In the cell body the polarity is as in
Perlpatus: that pole with the greatest amount of cytoplasm and containing the idiozome
mass is the distal pole (the one which in the dyaster stage of the last spermatogonic
division was in the equator of the cell). This polarity of the cell body is shown in Figs.
4, 5 and 8 ; I figured it also in a number of cases in my previous paper, but then errone-
ously supposed the idiozome mass to occupy that point where the spindle pole had previ-
ously been, whereas I am now able to determine positively that this pole is situated
directly opposite, namely, where the least amount of cytoplasm is situated. Now it would
appear in Euchistus, though not nearly so regularly as in Perlpatus, that it is the openings
of the U- or V-shaped bivalent chromosomes that are directed toward the distal pole of
the cell body (toward the pole where the idiozome mass is placed). Figs. 4, 5 and 8
show this for certain of the chromosomes, while other ones (as two in Fig. 5) may have
their openings in opposite directions. Thus in Euchistus there is more irregularity in the
positions of ^ the axes of the chromosomes, so that I have been unable to determine
whether it is, as in Peripatus, only particular ends of the univalent chromosomes which
unite with particular ends of others.

Throughout the growth period can be seen two kinds of liuin threads : (1) thicker
threads which connect the ends of the chromosomes, and (2) more delicate ones which
join chromatin granules with the nuclear membrane. Apparently, as I have shown for

OF THE GERM CELLS OF METAZOA. 159

Peripatus, the former, together with the linin contained in the chromosomes (axial
threads), together constitute a single continuous linin spirem in the nucleus.

Rest stage of the spermatocytes. A rest stage in the growth period preceding the
prophases of the maturation mitoses is well marked in Euchistus (Fig. 12 and Figs.
95-100 of my preceding paper) ; though I can confirm Paulmier's observation that such
a stage does not occur in Anasa. It is characterized by a huge true nucleolus, by a
rather diffuse and scattered distribution of the chromatin so that chromosomal boundaries
are practically indistinguishable, and by the diffuse arrangement of a great amount of
idiozome substance all around the nucleus, so that an idiozome mass with sharp outlines
is not present ; the idiozome mass in the synapsis stage (Figs. 4, 5, 8), on the contrary,
had a sharp and distinct outline. The bivalent chromatin nucleolus has now become
nearly rounded in form, rarely showed a dumbbell shape, so that its component parts are
very closely apposed. It lies peripheral, in contact with the nuclear membrane, while the
true nucleolus lies nearer the centre of the nucleus. Sometimes a much smaller rounded
body, staining like the chromatin nucleolus, is also found in the nucleus, but what its
origin is I have not been able to determine, for I have not found it in the spermatogonia,
though it might well escape detection there on account of its small size.

As to the terminology adopted by me in my former paper (1898) for the series of
stages of the growth period, which has been criticised by McClung (1900), the term
" metaphase " was, I grant, used by me incorrectly, for I used it for the commencement
of the anaphase, whereas it is really Strasburger's stage comparable to Flemming's " meta-
kinesis." However, the exact use of these terms was explained by me (1898, p. 20).
In the stages leading up to the resting spermatocyte I distinguished "early anaphase,"
" synapsis," " postsynapsis," and " telophase " as easily recognizable stages in the growth
period of Euchistus which need to be characterized by terms for purposes of description.
McClung (I. c.) considers the appearances of the synapsis stage as artefacts ; it is hardly
necessary to reply to this criticism, since in all Metozoa where the spermatogenesis has
been carefully examined, with the exception of certain Amphibia, the dense massing of
the chromosomes in the synapsis stage has been shown to be a perfectly normal phenome-
non. As to my use of "telaphase," Heidenhain's (1894, p. 524) definition is: "Unter
dem Namen Telekinesis beschreibe ich gewisse Bevvegungen des Kerns und des Mikrocen-
trums, welche gegen das Ende der Mitose hin stattfiuden. . . . Die zugehorigen Stadien
der Mitose bezeichne ich als Telophasen." Since Heidenhain employed it for the stage
just preceding rest in leucocytes, I was warranted in using it for the stage just before the
rest stage in the spermatocytes of Euchistus. It must be borne in mind, in the descrip-
tion of the changes of the growth period of the germ cells, that a peculiar stage, the
synapsis, occurs, not found elsewhere in mitosis, and that this stage modifies to greater or

160 MONTGOMERY A STUDY OP THE CHROMOSOMES

less extent, according to the object, the stages which precede and those which succeed.
It results from this that the stages of mitosis of the growth period cannot be exactly
compared with those of other cells, and hence the terms "anaphase" and ' ' telophase "
can here have a significance only approximately similar to that of other mitoses.

The mnfiirnfiijii divisions. In the early prophases the longitudinal splitting of the
chromosomes is well marked, clearer than in preceding stages (Figs. 13, 14). ^Each
chromosome is, as before the rest stage, clearly bivalent, formed of two longitudinally split
univalent chromosomes joined so as to make an angle together (Figs. 13-15), and at the
bend of the angle is a connecting linin thread. These forms of the bivalent chromo-
somes were clearly figured in my earlier paper, except that then I had overlooked the
longitudinal split. The chromosomes gradually become closer, shorter, with smoother out-
lines, the longitudinal split gradually becomes hidden, and the definitive chromosome with
the form of a dumbbell results (Figs. 16-18). In the definitive chromosome there is
usually no trace of the longitudinal split, except occasionally as a slight indentation at
the free end of a. univalent component. The constriction of the dumbbell marks the point
of union of two univalent chromosomes, which is effected by a linin band which gener-
ally never becomes quite hidden.

In the late prophases, just before the disappearance of the nuclear membrane, and
when the centrosome pairs have reached opposite poles of the nuclear surface, a remark-
able condition of the linin threads is found (Fig. 17) ; it was also shown in Figs. 152
and 153 of my earlier paper. The linin, previously in the form of fibres or strands,
now takes the form of chains of small globules quite as Van Beneden (1883) had
figured for Ascctris. I cannot explain this condition, but I have found it always at this
stage, and at this stage only.

In the first maturation division there are seven bivalent chromosomes and one bival-
ent chromatin nucleolus, and all these elements are divided transversely in metakiuesis,
being placed in the monaster stage so that their constrictions lie in the plane of the
equator. Fig. 18 shows a monaster stage with all these elements on lateral view, Fig.
19 on pole view ; this stage was accurately described by me in my former paper, so that
I have no additions to make to that description. Whole univalent chromosomes are
separated in the ensuing metakinesis, and the univalent components of the chromatin
nucleolus are also separated.

When the daughter chromosomes separate in the anaphase a constriction or indenta-
tion appears on them (Figs. 193-201 of my preceding paper). This I am now able to
prove, in agreement with Paulmier's observations on Anosa, is the reappearance of the
longitudinal split ; this indentation or constriction becomes placed in the equatorial plane
of the monaster stage of the second maturation division, so that the latter division divides

OF THE GERM CELLS OF METAZOA. 161

the chromosomes equatorially. In the anaphase of the first maturation division the con-
striction of the chromosomes generally has the appearance shown in Fig. 195 of my
former paper; while Fig. 196, which I then considered to represent the typical condition,
I now, from the study of more abundant material, find to be an unusual condition. That
is to say, the appearance of the chromosomes shown in Fig. 196 of my preceding paper
is really atypical, since in this case their constrictions appear at right angles to the long
axis of the spindle, whereas in most other cases the planes of these constrictions coincide
with planes passing through the long axis of the spindle. In this second maturation
division the chromatin nuclcolus is not always divided.

2. Euchistus tristiymus Say

Four testes of this species were studied.

In the rest stage of the spermatogonium there are two small chromatin nucleoli,
generally attached to the surface of the true nucleolus.

In the spermatogonic mitosis there are fourteen chromatin segments in the equatorial
plate (PL I, Fig. 20) ; the twelve larger, usually somewhat elongate ones are chromo-
somes, and the two smallest, rounded ones are chromatin nucleoli. All these elements
are halved in metakinesis.

In the synapsis stage the twelve chromosomes unite to form six bivalent chromo-
somes. The two chromatin nucleoli sometimes unite to form a bivalent one, which is
clearly dumbbell-shaped in earlier stages, but in the resting spermatocyte becomes
rounded and has a peripheral position (Fig. 22) ; or quite as frequently they remain
separate from one another during the growth period, and are seen to be of unequal vol-
umes (Fig. 21). The chromatin nucleoli in the growth period are rarely attached to the
true nucleolus.

In the first maturation division there are always six clearly bivalent, dumbbell-
shaped chromosomes and either one dumbbell-shaped bivalent chromatin nncleolus or,
apparently more frequently, two univalent chromatin nucleoli of more or less rounded
form and different volume (lateral view shown in Fig. 25, N. 2). Accordingly, on pole
views of the rnonaster stage there are seen either seven chromatin elements (Fig. 24),
which are six bivalent chromosomes and one bivalent chromatin nucleolus, or there are
eight, namely, six bivalent chromosomes and two univalent chromatin nucleoli (Fig. 23 :
in this figure one of the chromatin nucleoli can be distinguished by its smaller size, but
which of the remaining seven elements is the other chromatin nucleolus. is not easily dis-
cernible on pole views, since the larger of the two chromatin nucleoli has a diameter
equal to that of one of the smaller chromosomes).

All the six chromosomes are halved (by a reduction division) in metakinesis, so that

162 MONTGOMERY A STUDY OF THE CHROMOSOMES

in the monaster stage of the second division there are six univalent chromosomes, the
constrictions of which represent the reappearance of the longitudinal split (Fig. 20) ;
there are also in the same equatorial plate two non-constricted bodies of different volumes
which are not joined together. These are the chromatin nucleoli, which are regularly
halved in the first maturation metakinesis that is, they are the halves of univalent ones.
Thus the bodies marked N. 2 in Fig. 26 are the halves of those similarly marked in
Fig. 25.

The second maturation division is equatorial, and the spermatid receives six chro-
mosomes, arranged in an outer circle around a single central chromatin nucleolus.
Accordingly in this second division one chromatin nucleolus passes undivided into one
daughter cell (spermatid), the other undivided into the other daughter cell.

As in Euchistus variolarius, two follicles of the testis contain spermatocytes of a
much larger size than those in the four other follicles.

, 3. Podisus spinosus Dall.

Five testes of this species were studied.

In the spermatogonic rest stage there are two small chromatin nucleoli, of more or
less rounded form, attached to the surface of a true nucleolus.

In the spermatogonic monaster there are sixteen chromatin segments (PI. I, Fig.
27), two of which probably correspond to the chromatin nucleoli of the previous rest
stage.

In the synapsis the fourteen chromosomes unite to form seven bivalent chromosomes.
The two chromatin nucleoli also come together to make one bivalent one ; in the growth
period of the spermatocytes (Fig. 28) the chromatin nucleolus lies close to the nuclear
membrane, and to its inner surface the true nucleolus is regularly attached.

In the first maturation monaster there are eight chromatin elements, namely, seven .
chromosomes and one chromatin nucleolus, all bivalent and dumbbell-shaped on lateral
view ; the chromatin nucleolus has about the same volume as the smaller ones of the
chromosomes, and so cannot be distinguished from them with certainty.

4. Mormidea lugens Fabr.

Five testes were studied.

In the rest stage of the spermatogonia there are two chromatin nucleoli (PL I, Fig.
30, N. 2}, which may be equal or unequal in size ; they may be attached together, which
is apparently the general rule, or may be separated, and one or both of them may be
apposed to the true nucleolus.

In the spermatogonic monaster there are sixteen chromatin segments ; two of these

OK THE GERM CELLS OF METAZOA. 163

which are smaller than the others and more rounded are the chromatin nucleoli (Fig.
31, N. 3}.

In the synapsis the fourteen chromosomes unite to form seven bivalent ones, and the
two chromatin nucleoli to form one bivalent chromatin nucleolus. The latter is periph-
erally placed in the nucleus, and not attached to the true nucleolus (Fig. 32).

In a pole view of the monaster stage of the first maturation division are found eight
chromatin elements (Fig. 33) ; lateral view shows all are bivalent and dumbbell-shaped,
seven are chromosomes, and one easily recognizable by its much smaller size in the chro-
matin nucleolus (Fig. 33, N. 2).

5. Peribalm limbolaris Stal.

Two testes of this species were studied.

In the rest stage (PL I, Fig. 34) and prophases of the spermatogonia are found two
chromatin nucleoli of unequal size (N. 2], and sometimes apparently three ; they are
generally not in mutual contact, though they are often apposed to the true nucleoli, of
which there are frequently two or three.

In the spermatogonic monaster (Fig. 35) are sixteen chromatin segments, of which
the two smallest, rounded ones are the chromatin nucleoli ; the fourteen chromosomes are
notably elongated.

In the synapsis the fourteen chromosomes unite to form seven bivalent ones, and the
two chromatin nucleoli to form one bivalent chromatin nucleolus; in the rest stage of the
spermatocytes (Fig. 36) the chromatin nucleolus (N. 2) is usually rounded and peripher-
ally placed, and generally unattached to the relatively very large true nucleolus (some-
times there are two true nucleoli, as in Fig. 36, rarely three). In the rest stage of the
spermatocytes there is a smaller chromatin nucleolus in addition to the larger one already
described, and this smaller may correspond to the third of the chromatin nucleoli found
sometimes in the rest stage of the spermatogonia.

In the monaster stage of the first maturation division there are eight chromatin seg-
ments (Fig. 37), namely, seven bivalent, dumbbell-shaped chromosomes and one much
smaller bivalent, dumbbell-shaped chromatin nucleolus.

6. Cosmopepla carnifcx Fabr.

Five testes of this species were studied.

In the monaster stage of the spermatogonia (PL I, Fig. 38) are found eighteen
chromatin segments ; two of these are smaller than the others, and so by analogy with
other species of this family probably represent chromatin nucleoli (N. 2, Fig. 38) ; the
sixteen other segments are then true chromosomes.

1(54 MONTGOMERY A STUDY OF THE CHROMOSOMES

In the synapsis stage of the growth period the sixteen chromosomes unite to form
eight bivalent chromosomes, and the two chromatin nticleoli to form one bivalent chro-
matin nucleolus. The latter is, in the rest stage of the spermatocytes, rounded and
peripheral in position, and is not attached to the larger true nucleolus (Fig. 39) ; both
the nucleolus and the chromatin uucleolus may contain a large, clear vacuole, which in
the former is excentric.

Pole views of the monaster stage of the first maturation division show nine chroma-
tin elements (Fig. 41), and lateral views (Fig. 40) of the same stage show that all are
bivalent and dumbbell-shaped. The smallest of these elements is the chromatin nucleolus

(N. 2\

7. Nezara hilaris Say

Five testes of this form were studied.

There are in the rest stage and early prophases of the sperniatogonia two chromatin
nucleoli, which are comparatively large and usually more or less unequal in size (PI. I,
Figs. 42, 43, N. 2). They are generally peripheral in position and in mutual contact,
but usually are not apposed to the true nucleolus (-ZV).

In the spermatogonic monaster there are sixteen cliromatin segments (Fig. 44), of
which two can be always recognized by their small size and rounded form as the chroma-
tin nucleoli ; the fourteen chromosomes are generally elongated.

In the synapsis the two chromatin nucleoli unite to form one bivalent one, and
apparently also the fourteen chromosomes join to make seven bivalent chromosomes, but
I cannot state this with certainty. In the telophase of the spermatocytes (Fig. 45), the
chromatin nucleolus (N. 2) is peripherally placed and clearly bivalent, and usually not
in connection with the very large true nucleolus (-ZV), which is also peripheral.

In the testes examined (all from individuals secured in the month of September)
were stages only from the resting sperniatogonia to the telophase of the spermatocytes ; ;
all later stages in the spermatogenesis were absent, so that the number of the chromosomes
in the maturation divisions could not be determined.

The longitudinal split in the chromosomes during the growth period is unusually

distinct in this species.

8. Brochymena sp.

Three testes of this species were studied.

In the rest stage of the sperniatogonia (PI. I, Fig. 46) are two small chromatin
nucleoli (N. 2), which are peripheral in position, of nearly equal size, generally mutually
apposed, and seldom attached to the true nucleolus.

In the spermatogonic monaster stage (PL I, Fig. 47) are sixteen chromatin segments,
of which two are smaller and rounded and are the chromatin nucleoli.

OF THE GERM CELLS OF METAZOA. 165

In the synapsis stage the fourteen chromosomes unite to form seven bivalent ones,
and the two chromatin nucleoli to form one bivalent chromatin nucleolns. The latter in
the stages following the synapsis is rounded and peripheral in position (N. 2, Fig. 48,
PI. II), and only occasionally attached to the true nucleolus (N).

Pole views of the first maturation monaster (PI. II, Fig. 4i)) show eight chromatin
segments, of which one easily distinguishable from the others by its smaller size is the
chromatin nucleolus (N. 2). Lateral views of this stage show that all eight of these
elements are bivalent and dumbbell-shaped.

9, Perillus conftuens H.-S.

Two testes of this species were examined.

The rest stage of the spermatogonia (PI. II, Fig. 50) shows two small, rounded
chromatin nucleoli of unequal size, which are always attached together, and may be
either close to the nuclear membrane or apposed to the surface of a true nucleolus (N).

In the monaster stage of the spermatogonic divisions are sixteen chromatin segments
(Fig. 51). The fourteen largest are chromosomes, the two smallest are chromatin
nucleoli (N. 2) ; the latter are more minute than in the corresponding stage of any other
Pentatomid examined by me, and on account of their small size cannot always be seen
(i. e., in cases where they are closely apposed to the chromosomes).

In the synapsis stage the fourteen chromosomes unite to form seven bivalent ones
and the two chromatin nucleoli to make one bivalent chromatin nucleolus. The latter is
dumbbell-shaped in the earlier stages of the growth period, but in the rest stage (N. 2,
Fig. 52) becomes oval in outline, and it is then attached to the surface of the larger true
nucleolus (N. 2], the two occupying a more or less central position within the nucleus.

Pole views of the monaster stage of the first maturation division show eight chromatin
segments of varying diameter (Fig. 53) ; one of these, probably the smallest, is the
chromatiu nucleolus ; lateral views show that all these elements are bivalent and dumb-
bell-shaped.

10. Ccenus delius Say

Three testes of this species were studied.

In the rest stage of the spermatogonia there are two chromatin nucleoli with irreg-
ular outlines (PL II, Fig. 54, N. 2), and they are situated usually close together.

In the spermatogonic monaster stage (Fig. 55) there are fourteen chromatin seg-
ments, the two smallest of which are probably the chromatin nucleoli (N. 2*), leaving
twelve chromosomes.

In the synapsis stage the twelve chromosomes unite to form six bivalent ones. The

106 MONTi :oMKRY A STUDY OF THE CHROMOSOMES

two chromatin nuclcoli found in tlie spermatogonia also unite to form one bivalent
chroniiitin nucleolus ; this is clearly bilobed in the earlier stages, but more rounded in the
later stages of the growth period of the spermatocytes (the larger of the bodies designated
N. 2 in Figs. 57, 58, G3) ; there is attached to it usually a small true nucleolus (Fig. ">Si.
Besides this large bivalent chromatin mieleolus there is also found in the spermatocytes.
most clearly seen in the rest stage, another much smaller one, of rounded form (the
smaller of the bodies marked N. 2 in Figs. 57, 58, 63) ; this is almost always apposed to
one of the true nucleoli (N), of which there are generally two large ones besides the
small one attached to the large chromatin nucleolus ; quite frequently the small chro-
matin nucleplus lies between the large one and a large true nucleolus (Fig. G3). This
small chromatin nucleolus is difficult to see in the synapsis stage, when the chromosomes
stain deeply, and since I was also unable to find it in the monaster stage of the sperma-
togonia, I could not determine whether it is bivalent or univalent or what its earlier
history is. It might well be present, however, in the spermatogonia, but be there so
minute as to escape detection.

Pole views of the monaster stage of the first maturation division show sometimes only
seven chromatin segments (Fig. 62), and then these are six bivalent chromosomes and
the large bivalent chromatin nucleolus ; or they show eight segments (Figs. 59, 60), of
which the smallest is the small chromatin nucleolus of the growth period. That is to
say, in the equatorial plate there are always six bivalent chromosomes and the large
bivalent chromatin nucleolus, while the small chromatin nucleolus may be present or may
be absent. The lateral view of this stage given here (Fig. 61) shows seven large dumb-
bell-shaped elements, ol which six are chromosomes and one the bivalent chromatin
nucleolus though which one it would be hard to say, for all of these elements are of
approximately equal size and similar form ; while the smallest, eighth, element marked
N. 2 in this figure is the small chromatin nucleolus. When the latter persists into this
stage it appears to be halved in the following metakinesis.

II. Trichopepla semivittata Say

Four testes of this species were studied.

In the nucleus of the resting spermatogonium are seen clearly two rounded chro-
matin nucleoli (N. 2, Fig. 64, PL II), of different volumes, one or both frequently
apposed to a larger true nucleolus (N).

In the mouaster stage of the spermatogonia are found sixteen chromatin segments,
of which fourteen are elongate chromosomes, and two which are smaller and rounded are
the chromatin nucleoli (N. 2, Fig. 65), which here, as in the preceding rest stage, are
unequal in size.

OF THE GERM CELLS OF METAZOA. 167

In the following synapsis stage the fourteen chromosomes join to form seven bivalent
ones. The two chromatin nucleoli likewise unite to form one bivalent one, of which the
two components are unequal in size (Fig. 66, N. 2\ In the telophase and rest stage of
the spermatocytes the chromatin nucleolus loses its earlier bipartite form and becomes
rounded (the larger of the bodies marked N. 2 in Fig. 67), and only occasionally is it
apposed to the larger true nucleolus (N). Sometimes the two chromatin nucleoli derived
from the spermatogonia do not unite together, but remain separated. During the growth
period, its later stages at least, can be seen in each nucleus three or four much smaller,
rounded bodies, which stain like the chromatin nucleoli ; some of them are often attached
to the surface of the true nucleolus (Fig. 67, the three smaller bodies designated N.' 2).
There are certainly three of them and in some nuclei apparently four. I was unable to
determine with certainty these small chromatin nucleoli in the rest and division stages of
the spermatogonia, though they might well be present there, but escape observation on
account of their minuteness.

The monaster stage of the first maturation division (Figs. 68, 69) shows eight larger,
bivalent, dumbbell-shaped chromatin segments, of which seven are chromosomes and one
the large chromatin nucleolus (N. 2 of the figures). Of the seven chromosomes one is
always longer and more voluminous than the others (Figs. 68, 69), and is probably the
derivative of the two largest chromosomes found in the spermatogonic divisions (Fig. 65).
Besides these eight large elements of the monaster stage of the reduction division there
may be seen on pole view usually one (Fig. 68), sometimes two much smaller granules,
which evidently represent the small chromatin nucleoli found in the growth period.

SCUTELLARIID.E.

12. Eurygastcr alternatus Say

Three testes of this form were studied from individuals taken in July and August.
Each testis was filled with spermatocytes and spermatids, but contained no spermatogonia.

In the spermatocyte in the rest stage one bilobed and hence probably bivalent
chromatin nucleolus (N. -2, Fig. 70, PI. II), which is peripheral in position and separated
from the usually smaller true nucleolus (-^V). Sometimes the two components of this
chromatin nucleolus do not join together in the synapsis but remain separated through
the growth period.

In the monaster stage of the first maturation division (Fig. 71, pole view) are found
seven dumbbell-shaped (and hence probably bivalent) chromatin segments, of which the
smallest is undoubtedly the chromatin nucleolus (N. #), so that here there would be six
bivalent chromosomes.

168 MONTGOMERY A STUDY OF THE CHROMOSOMES

COREIDJE.

13. Anasa tristis De G.

Twenty-one testes of this species were studied.

In regard to the chromosomal numbers my observations confirm those of Paulmier.

In the rest stage of the spermatogonium (PI. II, Figs. 72, 73) there are two chro-
matin nucleoli (N. #), which are much smaller than the true nucleoli (N) to which they
are generally apposed. They have definite irregularly rounded or oval outlines as
examined with Hermann's saffranine-gentian violet stain, and are not " hazy " or " indefi-
nite " as Paulmier (1899) described. Both may be attached to the same nucleolus, or they
may be joined to separate nucleoli. Sometimes each one may separate into two pieces (as
is the case with one in Fig. 72). They are best seen on iron haematoxylin preparations
so strongly destaiued that the chromatin reticulum does not appear.

In the monaster stage of the spermatogonia (Fig. 74) are twenty-two chromatin seg-
ments, namely, twenty larger chromosomes and two much smaller chromatin nucleol 1
(N. *).

In the synapsis stage the twenty chromosomes unite to form ten bivalent ones, and
the two chromatin nucleoli to form one bivalent one. The latter is clearly bipartite in
the synapsis, but later shows an oval outline (Fig. 75) ; it is peripheral in position, often
contains a central clearer vacuole as in the Pentatomidce, and is as a rule separated from
the true nucleolus (^V).

In the monaster stage of the first maturation division (pole view, Fig. 76) are found
eleven bivalent, dumbbell-shaped chromatin segments, of which the central, smallest one
is the chromatin nucleolus (N. 2). I am able to confirm Paulmier's (1899) account of
the two maturation divisions.

14. Anasa armigera Say

One testis of this species was studied.

The spermatogeuesis seems to be very similar to that of the preceding species, but
as I had no preparation stained with saffranine-gentian violet I was unable to determine
the relations of the chromatin nucleoli in the rest stage of the spermatogonia.

Fig. 77, PI. II, shows a pole view of a monaster stage of the spermatogonia, with
twenty chromosomes and two chromatin nucleoli (N. 2}; it is very similar to the corre-
sponding stage in Anasa tristis (Fig. 74).

In the synapsis are formed ten bivalent chromosomes and one bivalent chromatin
nucleolus.

In the monaster of the first maturation division (Fig. 78) are ten bivalent chroino-

OF THE GERM CELLS OF METAZOA. 169

somes and one bivalent chromatin nucleolus; Fig. 78 is a pole view, but in it those
elements which appear dumbbell-shaped are seen from the side.

15. Anasa sp.

Of this undetermined species, which was collected for me at Berryessa in California
I examined nine testes.

The resting spermatogonium shows two chromatin nucleoli (PI. II, Fig. 79, N. 2)
which are comparatively large and rather loose in texture, generally irregular in outline,
occasionally attached to the true nucleolus ( N ), and more or less central in position.

In the monaster stage of the spermatogonia (Fig. 80) are twenty-two chromatin
segments, namely, twenty larger chromosomes and two smaller chromatin nucleoli (Of. 2).

In the synapsis the chromosomes unite to form ten bivalent ones, and the chromatin
nucleoli to form one bivalent one. In the rest stage of the spermatocytes (Fig. 81) the
chromatin nucleolus (JV. ) is seen to be somewhat elongate in form, is peripheral in
position, and not attached to the true nucleolus (-ZV).

In the monaster stage of the first maturation division are eleven bivalent elements,
of which the smallest is the chromatin nucleolus (N. 2, Fig. 82) ; in this figure we do
not have strictly pole views of all the chromosomes.

Fig. 83 shows four of the bivalent chromosomes on lateral view, in a paratongential
section of a cell in the stage of the first maturation monaster. It is given here because
it is the clearest case I have noticed in any Hemipteron of the quadripartite nature of
these chromosomes, for while the transverse split may generally be seen at this stage, the
longitudinal split is generally hidden. The poles of the spindle (not in the plane of
this section, but seen in the next one to it) are situated at the upper and lower portions of
the figure respectively ; and it is hardly necessary to add that the first maturation division
coincides with the plane of the transverse split, the second with the plane of the longi-
tudinal split.

16. Metapodius terminalis Dall.

Eleven testes were studied of this species, which is very favorable on account of the
large size of the cells ; one should examine testes from individuals taken in June or early
July, before the time of copulation.

In the rest stage of the spermatogonia (PI. II, Fig. 84) are two chromatin nucleoli
(N. ) of very small size and smooth outlines, generally close together on the surface of
a true nucleolus (N).

In the spermatogonic monaster (Fig. 85) are twenty-two chromatin segments, of
which the two smallest are chromatin nucleoli (N. 2) and easily recognizable.

170 MONTGOMERY A STUDY OF THE CHROMOSOMES

In the synapsis the twenty chromosomes unite to form ten bivalent ones, and the
two chromatin nucleoli to form one bivalent one. The latter is in later stages of the
growth period bilobed (N. 2, Fig. 86), is peripheral in position and not opposed to the
larger true nucleolus (N).

In the monaster stage of the first maturation division are eleven chromatin segments
(Fig. 87), of which the smallest, centrally placed one is the chromatin nucleolus (N. 2) ;
all these elements are bivalent and on lateral view they all show the dumbbell-shape.

17. Chariesterus anlennator Fabr.

Two testes of this form were examined.

In the rest stage of the spermatogonia I could not be certain of the presence of chro-
matin nucleoli, for my preparations were not very well stained to demonstrate them.
There were also no spermatogonic monasters favorable enough for determining the num-
ber of chromosomes.

In the synapsis stage there is a bivalent chromatin nucleolus, but sometimes its com-
ponent parts are widely separated.

In the telophase of the spermatocytes (PI. II, Fig. 88) the chromatin nucleolus (N.
#) is peripheral in position, sometimes its two univalent components still separated (but
that is not the case in Fig. 88). The true nucleolus (N) is sometimes central, sometimes
peripheral in position, and occasionally it is apposed to the chromatin nucleolus.

In the monaster of the first maturation division (Fig. 89, lateral view ; Fig. 90, pole
view) are found thirteen chromatin segments, of which the smallest, centrally placed one
is the bivalent chromatin nucleolus (N. 2). Of the twelve chromosomes at least eleven
would seem to be bivalent (having the characteristic dumbbell-shape) ; but in the lateral
view here given (Fig. 89), it will be noted that the chromosome nearest the left-hand
side does not appear dumbbell-shaped. This may be a bivalent one seen obliquely, or it
may be a univalent one ; which is the case I cannot determine, since there were few
satisfactory lateral views on the preparations and since the number of chromosomes in
the spermatogouia could not be determined.

18. Alydus pilosulus H. 8.

Four testes of this species were studied.

The chromatin nucleoli in the rest stage of the spermatogonia (N. 2, Fig. 91, PL
II) are two in number and rounded ; they are very small, usually close together, and
may be or not be attached to the true uucleolus (N).

In the spermatogonic monaster stage (Fig. 92) are fourteen chromatin segments, l\v>

OF THE GERM CELLS OF METAZOA. 171

of which, easily distinguishable from the others by their small size, are chroinatin
nucleoli (JV. %).

In the synapsis stage the twelve chromosomes unite to form six bivalent ones, and
the two chromatin nucleoli to form one bivalent one. In the late stages of the growth
period (Figs. 93, 94) the chromatin nucleolus (N. #) is rounded and peripheral in posi-
tion, and usually apposed to the true nucleolus ; even when they are separated the latter
is usually peripheral (N, Fig. 94) an unusual position for it in spermatocytes of
Hemiptera.

In the monaster stage of the first maturation division (Fig. 95) are seven elements,
namely, six bivalent chromosomes and one bivalent chromatin nucleolus (the smallest of
the seven elements, N. 2] ; all these are dumbbell-shaped on lateral view, and though
Fig. 95 is a pole view of the spindle two of its chromosomes are seen from the side.

19. Alydus eurinus Say

One testis of this species was studied.

In the rest stage of the spermatogonia I could not determine chromatin nucleoli,
probably on account of their small size.

Numerous monaster stages of spermatogonia were examined, and all showed thirteen
chromatin elements (PI. Ill, Fig. 96) ; two of these which are readily recognizable from
the others by their minute size are chromatin nucleoli (the two small granules shown in
Fig. 90) ; the eleven large elements are chromosomes, and have mostly an elongated
form.

In the synapsis the two chromatin nucleoli unite to form one bivalent one, which in
the telophase of the spermatocytes (N. 2, Fig. 97) is relatively small, peripheral in posi-
tion, and quite frequently apposed to the larger true nucleolus (N). Of the eleven uni-
valent chromosomes derived from the spermatogonium, ten unite to form five bivalent
pairs, while one (the eleventh) does not unite with any of the others but remains uni-
valent.

In the first maturation division are found seven chromatin elements (Fig. 98, pole
view) ; the smallest of these is bivalent, dumbbell-shaped, and is the chromatin nucleolus
(N. 2} the six larger elements are chromosomes. Now a careful study of numerous
monaster stages seen on lateral view shows that only five of these chromosomes are dumb-
bell-shaped, and so bivalent on analogy with what is known for the other Hemiptera ;
while one of them is never dumbbell-shaped, approximately half the volume of the
others, and is univalent. Fig. 99 is a lateral view of the spindle of the first maturation
division, showing three of the five dumbbell-shaped chromosomes and (most to the right)
the univalent chromosome. In all cases where the chromosome plates of this stage can

172 MONTGOMERY A STUDY OF THE CHROMOSOMES

be seen on lateral view, one chromosome is always found to be of about half the size of
the others and not dumbbell-shaped ; on pole views this chromosome can be distin-
guished by its lesser depth.

So in Alydus eurinus there is an uneven number of chromosomes in the spermato-
gonia, namely, eleven ; the reduction in number is effected then in the synapsis by ten
combining to form five bivalent ones, while one remains uiiivalent and uncombined.
because there is no mate with which it can unite.

In the monaster of the second maturation division there are either six or seven
chromatin elements. In Fig. 100 of this stage are shown seven, of which the smallest is
probably the chromatin nucleolus, five are halves of the originally bivalent chromosomes
and one probably the half of the originally univalent chromosome.

In the spermatid we find either six (Fig. 102) or five (Fig. 101) chromatin elements
of approximately equal volume. Now these elements are too large to be derivatives of
the chromatin nucleolus of the spermatocyte of the first order (N, 2, Fig. 98), so that the
five or six elements of the spermatids would not seem to represent portions of this chro-
matin nucleolus ; very probably the latter is so small in the spermatids or generally so
closely applied to the surface of one of the chromosomes that it escapes observation. If
we then eliminate the possibility of any of the elements shown in the spermatids (Figs.
101, 102) representing chromatin nucleoli or their derivatives, then we must conclude that
the five or six elements here are chromosomes. But why is their number sometimes five,
in other cases six ? Now we know that in all other Hemiptera in which attention has
been given to this point that each spermatid receives one-quarter of each of the bivalent
chromosomes present in the spermatocyte of the first order. Accordingly it would be
probable by analogy that in Alydus eurinus the spermatid receives one-quarter of each of
the original five bivalent chromosomes. Then in the case of Fig. 101 all five elements
would be such derivatives ; in Fig. 102, five of the six elements. The sixth element of
Fig. 102 is then probably the original univalent chromosomes of the first maturation
division, which in either the first or the second maturation division could not have been
divided, but must have passed undivided into one of the daughter cells ; this would
explain why sometimes there are only five, sometimes six chromosomes in the spermatid,
for, as I have explained, none of the elements of Figs. 101 and 102 can be regarded as
chromatin nucleoli.

Of course the preceding is only an attempt at a right interpretation ; I have not been
able to follow the univalent chromosome with precision in regard to its behavior in the

maturation divisions.

20. Oorizus lateralis Say

Four testes of this species were studied.

I could not determine whether there are chromatin nucleoli in the rest stage of the

OK THE GERM CELLS OF METAZOA. 173

spermatogonia, and none of the cases of spermatogonic monastery in my preparations were
sufficiently favorable to allow accurate counting of the chromatin elements.

In the growth period of the spermatocytes a comparatively small, bivalent chromatin
nucleolus, which in the rest stage (N. 2, Fig. 103, PI. Ill) has a peripheral position and
rounded form and is not apposed to the true nucleolus (N)- Besides this chromatin
nucleolus one or two much smaller, rounded ones can sometimes be seen in the nuclei of
the resting spermatocytes, and these stain like the large one with the double stain of
Hermann.

In the nionaster stage of the first maturation division are always found at least seven
chromatin elements ; when there are eight the eighth is a small granule (Fig. 105, the
smaller of the bodies marked N. 2), and this small element, which frequently cannot be
seen at this stage, probably represents one of the minute chromatin uucleoli of the
growth period. Of the seven larger elements the smallest, centrally placed one is the
bivalent chromatin nucleolus (N. 2, Fig. 104, and the larger of the elements marked
N. 2 in Fig. 105) ; this chromatin nucleolus often has its component halves separated
(except for a joining linin band) before the period of metakiuesis (Fig. 106). The
remaining six elements are chromosomes, and of them four are of approximately equal
volume, while one is always much larger and one always much smaller than these four
(pole views Figs. 104, 105, lateral view Fig. 106). The five largest chromosomes are
clearly dumbbell-shaped on lateral view, and accordingly by analogy with the corre-
sponding elements of other Hemiptera may be considered bivalent, even though the
number in the spermatogonia was not determined. But the smallest chromosome puzzled
me at first with regard to its valence, for it is not more than half the volume of the other
five, and sometimes it does not appear dumbbell-shaped, so that I considered the possi-
bility of its being uriivalent ; but a careful study of it in numerous cells of the first
maturation division resulted in showing in a number of clear cases that it is transversely
constricted even before it becomes arranged in the plane of the equator, so that there can
hardly be a doubt as to its being bivalent. Fig. 106 shows such a case in an oblique
lateral view of the spindle before the chromosomes have become arranged all in one
plane, with a well-marked constriction of the smallest chromosome.

The first maturation division is a reduction division and always halves the six chro-
mosomes and the bivalent chromatin nucleolus.

21. Harmostes reflexulus Stal.

Thirteen testes of this species were studied.

In the rest stage of the spermatogonia there are two rounded chromatin nucleoli
2, Fig. 107, PL III), of which one or both may be apposed to the true nucleolus (N).

174 MONTGOMERY A STUDY OF THE CHROMOSOMES

Iii the monaster stage of the spcrmatogonic divisions can always be counted thirteen
chromatin segments on favorable pole views i. e., in such cases where these elements
can be seen all in one plane, and where they are not too closely apposed to one another
(Figs. 108-110). Two of these elements are always distinguishable from the others by
their smaller size and rounded shape, and these are the chromatin nucleoli (N. 2) ; they
may lie close together (Fig. 110), but more usually are more separated in position (Figs.
108, 109). The remaining eleven elements, which are of large size and elongate form,
are chromosomes. There can be no doubt that this is the actual number of these chro-
mosomes, for no exceptions to it were found, and in fourteen clear cases from, four different
testes the number eleven was obtained with great clearness ; these chromosomes are larger
than the spermatogonic chromosomes of any other Hemipteron examined. Sometimes
one or more of the chromosomes may show a slight transverse constriction, but this is not
a constant appearance.

In the synapsis the two chromatin nucleoli derived from the spermatogonium unite
to form one bivalent one ; in the rest stage (which is very complete) of the spermatocyte
it is elongate (_ZV. 2, Fig. Ill), peripheral in position and not attached to the true
nucleolus ; the latter is larger (N, Fig. Ill), frequently peripheral in position, and
sometimes two true nucleoli are present.

During the synapsis stage ten of the eleven chromosomes join to form five bivalent
chromosomes, while the eleventh remains univalent, as will become evident from the
following description :

In the first maturation division are found either seven chromatin elements or eight
chromatin elements ; these two conditions may be described successively.

When there are seven elements (Fig. 112 pole view of the monaster stage, Fig. 113
lateral view) one may always be distinguished by its smaller size and central position,
and by its history from the rest stage of the spermatocyte to the stage under discussion
this is found to be the chromatin nucleolus (.2V. 2) ; this we have already learned to be
bivalent, and in Fig. 113 its univalent components are seen to be separating. The six
larger elements are chromosomes. Five of these, as Fig. 113 shows, are clearly dumbbell-
shaped and bivalent. The sixth, however (), never shows a dumbbell shape before the
metakinesis, but is always distinguishable from the others 'by its oval form. From these
appearances we must conclude that this sixth chromosome is univalent represents the
odd, eleventh, chromosome of the spermatogonic monaster stage, which had no mate with
which to unite during the following synapsis stage. It was in this species that I first
found spermatogonia with an uneven number of chromosomes, so that I first concluded
they must be abnormal cases, for heretofore in all objects the spermatogonic (normal)
number has been described as an even one ; I immediately sectioned testes of other

OF THE GERM CELLS OF METAZOA. 175

individuals to determine this point, but, as has been already stated, in the four of the
thirteen testes which contained spermatogonic divisions exactly eleven chromosomes
were found to be always present. Obviously all of the eleven chromosomes cannot unite
into pairs during the synapsis, one must remain unmated, and this must be necessarily
that one of the first maturation division which does not appear bipartite.

Now in those cases where there are eight chromatin elements present in the sperma-
tocytes the question becomes more complicated (Figs. 114-116). Here, as in the cases
where there are seven elements, one is central in position and distinguishable from the
others by its smaller volume, namely, the bivalent chromatin nucleolus (N. 2, Figs.
114-116, in the last figure its univalent components separated) ; the remaining seven
elements are then chromosomes. As the lateral view, Fig. 116, of the spindle shows, four
of these are bivalent (the ones not marked by lettering). One (x) is oval in form, show-
ing no constriction or splitting, and so is probably comparable to the univalent chromo-
some of those cells which contain but seven chromatin elements (i. c., to x in Fig. 113.)
There then remain the two elements marked a in Fig. 116, and each of these I conclude
must be a univalent chromosome, which in cases where there are only seven chromatin
elements in the spindle would have combined with the other to form one bivalent chromo-
some ; if this be so, then the transverse constriction of the left-hand chromosome marked
a in Fig. 116 would not be the line of separation between two univalent chromosomes.
Another reason for looking upon these two chromosomes as univalent, is because they are
of approximately the same volume as the chromosome marked x, which we have shown
to be univalent by comparison with the chromosome x of Fig. 113. But there is a still
better reason for considering the elements a of Fig. 116 to be univalent chromosomes.
A pole view of a corresponding stage with eight chromatin elements shows the seven
chromosomes frequently equidistant from one another (as in Fig. 115). But often we find
on pole view two of the seven chromosomes close together and connected by a band of
linin (a, Fig. 114) ; the two together constitute a virtual bivalent chromosome, which,
however, differs from the other bivalent ones in having its long axis parallel to the plane
of the equator of the spindle. Let the band of linin which connects these chromosomes
a become stretched out, and as a result we would have a bivalent chromosome lying
parallel to the plane of the equator, and with its univalent halves widely separated i. e.,
the condition that maintains for the chromosomes a of Fig. 116.

To summarize, we find two conditions in the spermatocytes : (1) there are seven
chromatin elements, namely, one bivalent chromatin nucleolus, five bivalent chromo-
somes, and one univalent chromosome > and (2) eight chromatin elements, namely, one
bivalent chromatin nucleolus, four bivalent chromosomes, one univalent chromosome
(corresponding to that of condition 1), and two other univalent chromosomes (which

176

MONTGOMERY A STUDY OF THE CHROMOSOMES

together would correspond to the fifth bivalent chromosome of condition 1). To deter-
mine which of these conditions is the more usual, I counted the number of chromatin
elements seen on pole views of the monaster stage of the first maturation division. These
counts were made on spermatocytes from five different testes, and may be condensed into
the following table :

PREPAnATION NO.

EIGHT ELEMENTS.

SEVEN ELEMENTS.

86

4

33

238

6

12

408

2

1

410

1

9

356

23

Total =

13

78

Thus those spermatocytes with seven chromatin elements would seem to be the more
frequent condition. In both cases there is one univalent chromosome, which represents
the odd chromosome of the spermatogonia ; but why in cells of the second condition two
chromosomes should remain separated instead of combining to form a bivalent one, as they
do in the first condition, I cannot explain, unless perhaps the presence of the odd uni-
valent chromosome may in some way disturb the union into pairs of the ten other chro-
mosomes during the synapsis.

In those cases where there are seven chromatin elements in the equator of the first
maturation spindle, the metakinesis results in the division of all the elements; this is a
reduction (transverse) division of the bivalent chromatin nucleolus and of the five bival-
ent chromosomes, but in what plane the univalent chromosome divides could not be de-
termined. Only one case was seen where the univalent chromosome was left undivided
in the equator after the daughter elements of the six other elements had reached opposite
poles of the spindle. Thus it would seem that in this division, in the cases where there
are seven elements present, all the elements become divided ; how it is in the cases where
there are eight elements could not be determined. In the spermatic! are found either six
chromosomes (Fig. 117) and one chromatin nucleolus (N. 2), or five chromosomes and
one chromatin nucleolus. This would show that the chromatin nucleolus and five chro-
mosomes (the derivatives of the original five bivalent ones) divide in the second matura-
tion division, but that the sixth chromosome, the derivative of the originally univalent
one, does not divide but passes undivided into 'one of the two spermatids. Thus the
valence of the seven elements in these generations would be : first spermatocyte, one
bivalent chromatin nucleolus, five bivalent chromosomes, one univalent chromosome ;

OP THE GERM CELLS OF METAZOA. 177

second spermatocyte, one univalent chromatin nucleolus, five uuivalent chromosomes, one
semivalent chromosome ; spermatid, one semivalent chromatin nucleolus and either five
or six semivalent chromosomes.

22. Protenor bclfragei Hagl.

Five testes of this exceedingly interesting species were studied from individuals that
had just completed their last ekdysis.

In the rest stage of the spermatogonia (PI. Ill, Fig. 118) are two rounded chroma-
tin nucleoli which are usually attached to the surface of a much larger true nucleolus (N).

Pole views of the monaster stage of the spermatogonic mitosis show with great dis-
tinctness exactly thirteen chromatin elements (Figs. 119-123). This number was found
in thirteen cells of one testis, in ahout sixteen cells of a second, in six cells of a third,
and in two cells of a fourth these being all the favorable cases found, and all these
testes had been fixed with Flemming's fluid (the stronger mixture). The fifth testis
sectioned had been fixed in picro-acetic acid, and in it the number of chromosomes could
not be counted because of the swelling action which this reagent exerts upon the chroma-
tin. These chromosomes are unusually large and on suitable preparations can be counted
with exactness. In only two of the cells in which they were counted was there observed a
fourteenth element; this was a minute granule (/, Fig. 121), which, on account of its
being present so rarely in these monaster stages and on account of there being no
element to represent it in the later history of the spermatogenesis, need not be taken into
account ; it seems to be very inconstant, and might possibly represent either a portion of
chromatin which had become separated from one of the chromosomes, or a chromatin
iiucleolus transmitted from some distant parent and now nearly reaching disappearance.

Which two of the elements in the spermatogonic monaster represent the chromatin
nucleoli of the previous rest stage I am unable to determine, but that two of them do
represent these bodies there can be no doubt from what has been determined for the suc-
ceeding stages; judging by analogy with the case in all other Coreidce examined, they
would probably be the two smallest elements. Now Figs. 119-123 show what is to be
seen very distinctly in all cases, namely, that there are three chromatin elements much
larger than the ten remaining. One of these three, that designated x in Figs. 119-123,
imposes by its relatively very large volume ; this has in most cases the form shown in
Figs. 120 and 121, but in a few cases it was noticeably elongated, as in Figs. 122 and
123. The last figure shows it to have a transverse constriction around the middle ; and
this case, together with the fact of its great volume, would show it to be equal potentially
to at last two chromosomes ; for purposes of description we shall call this the " chromo-
some x." The two other chromosomes, which can always be recognized by their relatively

178 MONTGOMERY A STUDY OF THE CHROMOSOMES

large volume, are those designated by the letter k in Figs. 119-123; these two are of
approximately equal volume, and each has about half the volume of the chromosome x.

There are accordingly present in the spermatogonic monaster thirteen chromatin
elements, of which two (probably the smallest) represent the chromatin nucleoli ; of the
eleven chromosomes, three are much larger than the others, namely, the one marked x
and the two marked k in the Figs. 119-123. In the metakinesis all these elements are
halved longitudinally.

In the following synapsis stage we find a small chromatin nucleolus composed of two
parts, which in every way is comparable to the bivalent chromatin nucleolus of the growth
period of other Coreidce ; this is marked N. 2 in Figs. 124, 129, 130. This chromatin
nucleolus is peripheral in position, and only occasionally has a true nucleolus apposed to
it (Fig. 130). Generally its two univalent halves are not closely apposed but more or
less separated, often widely separated (N. 2, Fig. 131), but the two always come close
together to form a dumbbell-shaped, bivalent body before the monaster stage of the first
maturation division. Certainly its two components must represent the two univalent
chromatin nucleoli of the rest stage of the spermatogonia (N. 2, Fig. 118).

During the synapsis stage also ten out of the eleven chromosomes derived from the
spermatogonium combine to form five bivalent chromosomes, as will be shown in treating
of the maturation divisions. The odd one of the eleven chromosomes does not combine
with any other during the synapsis stage, and this is the largest of the chromosomes of
the spermatogonium, namely, the chromosome x. This element has a remarkable history
in the growth period. Through the whole growth period it acts like a chromatin nucle-
olus in preserving a compact form and in continuing to take the safframne stain with the
use of the double stain of Hermann, while the other chromosomes take the violet stain.
It will be remembered that this chromosome x becomes distinguishable first in the sperma-
togonic mitoses (Figs. 119-123), while in the preceding spermatogonic rest stage it
cannot be distinguished, for then it takes the violet stain like the other chromosomes and
takes part in the formation of the nuclear riticulum just as they do ; accordingly it can
be concluded that it commences to behave differently from the other chromosomes at the
beginning of the growth period of the spermatocyte. In the early synapsis (Fig. 124)
it has the same general shape as in the spermatogonic monaster stage (compare the ele-
ment marked x in Fig. 124 with the corresponding one in Figs. 120, 121), but it has
greatly increased in volume, as a comparison of these figures show, since it will be recalled
that the chromosome x of the spermatccytes is a half of the chromosome x of the sperma-
togonia. Later in the growth period the chromosome x elongates into the form of a bent
rod (Figs. 125130), which usually lies close to the nuclear membrane (in this point
also resembling a chromatin nucleolus); throughout the growth period it keeps its com-

OF THE GERM CELLS OF .METAZOA. 179

pact structure and smooth outline. When it is beginning to elongate a faintly-marked
clear line can be seen in its long axis (Figs. 125, 129), and this is evidently a longitudi-
nal split, comparable to that of the bivalent chromosomes ; this split cannot be seen in
the telophase nor at any period of the maturation divisions. About coincidently
appears a transverse split; this maybe a simple annular constriction, or a clear con-
necting bridge of linin (as in Fig. 128). This transverse constriction, pointing to a
bipartite nature, would show that the chromosome x is bivalent ; but it must have been
already bivalent in the spermatogonia (where also a transverse constriction can some-
times be seen, Fig. 123), for it does not unite with any other chromosome in the sperma-
tocytes. I can find no other explanation for its occasional bipartite appearance during
the synapsis.

In the later period of the synapsis stage, in the telophase, and in the early prophases
of the first maturation division the chromosome x undergoes considerable changes in form.
The slightly bent rod (Figs. 125, 129) of the early synapsis bends at its middle point,
where the transverse constriction was apparent, into the form of~a U or V (Figs. 126,
127, 130), or even the form of an S (Fig. 1275). The end result of these bendiugs seems
always to be a horseshoe-shape (Fig. 130) or a nearly closed ring (Fig. 127c). From the
early synapsis stage until the beginning of the prophases of the first maturation mitosis
a true nucleolus of varying form is attached to the surface of the chromosome x (N,
Figs. 124, 125, 127130); occasionally this true nucleolus may be separated into two or
three parts, all of them attached to the chromosome. In the prophases of the first
maturation division the nucleolus becomes detached from the chromosome, rapidly
decreases in size, and becomes lost before the nuclear membrane disappears.

In the early prophases of the first maturation division (Figs. 131, 132) are found
the following elements: (1) five bivalent chromosomes, of which all five are shown in
Fig. 132, only four in Fig. 131 (all these seen on lateral view); all these show at this
stage a well-marked longitudinal split (often of circular or oval outline) and a trans-
verse split (which marks the point of union of two univalent chromosomes); the mode
of formation of these elements is very similar to that described by Paulmier (1899) for
Anasa. (2) The bivalent chromatin nucleolus, the two parts of which may be in close
contact (N. 2, Fig. 132) or may still be widely separated (N. 2, Fig. 131). And (3)
the large chromosome x, the largest of all the elements, which now has decreased some-
what in volume owing to the greater condensation of its substance ; seen from the side,
it gives the appearance of a thick horseshoe or a nearly closed ring (x, Fig. 131). Of
the five bivalent chromosomes, one is always much larger than the others (K. 2, Figs.
131, 132), and this was evidently formed by the union of the two large chromosomes
A^of the spermatogonic mitoses (Figs. 119123).

180 MONTGOMERY A STUDY OF THE CHROMOSOMES

In the later prophases (Figs. 133, 134) the five bivalent chromosomes condense into
the form of dumbbells or sometimes of rings, the large chromosome K. 2 (Fig. 133)
passing through these stages more slowly than the others, so that it often retains loose
texture and roughened outlines after the four others have become compact with smooth
outlines. The bivalent chromatin nucleolus now has its univalent components generally
iu rather close apposition (N. 2, Figs. 133, 134). The chromosome x is very compact
in structure, and when seen from the side has squarish form (x, Fig. 133), an indentation
at one end of which marks the point of apposition of the ends of the primitive horse-
shoe form which latter in some cases (Fig. 134) may still be seen at this late stage.
Usually the chromosome x is longer than broad, and the clear line sometimes found in its
long axis does not then represent the primitive longitudinal split, of which there seems
no trace at this stage, but the space separating the two arms of the horseshoe.

In the inonaster stage of the first maturation mitosis (Figs. 135-137) there are
accordingly seven elements (in the cell from which Fig. 136 was drawn, one of the
bivalent chromosomes lay out of the plane of the section). These are the bivalent
chromatin nucleolus (N. &), the smallest of all ; the chromosome x (x), the largest of
all ; and five bivalent chromosomes, of which one is almost always recognizable by its
greater volume (K. 2), and this is the bivalent chromosome formed by the synapsis of
the two larger chromosomes A" of the spermatogonia. The five bivalent chromosomes
and the chromatin nucleolus become divided transversely in the metakinesis (reduction
division). The chromosome z (for successive stages in its division, Figs. 13G, 138),
which has its long axis coinciding with the plane of the equator, becomes divided into two
along its median axis. This would appear at first sight to be a longitudinal (equational)
division; but it is not, for we have learned that this peculiar chromosome had first t In-
form of a straight rod, which then bent at its middle point into a U or V, then the arms
of the U or V laid themselves parallel to and close together, so that a division along the
median axis results now in the separation of these arms, and is accordingly a reduction
division.

A view of the second spermatocyte, before its chromatin elements have definitely
arranged themselves in the plane of the equator of the spindle, shows also seven elements
(Fig. 139) : one univaleut chromatin nucleolus (N. 2) ; one chromosome larger than the
others, a half of the original chromosome x (x i); and five univalent chromosomes, one
(A") larger than the others and directly comparable to one of the large chromosomes A of
the spermatogonia. In the metakinesis following (Fig. 140) the five univalent chromo-
somes are divided equationally, and the univalent chromatin nucleolus (N. %) is also
divided (but in what plane was not determined). The chromosome x \, however, never
becomes divided in this mitosis, but passes undivided into one of the daughter cells (Fig.

OF THE GERM CELLS OF METAZOA. 181

140); and in the dyaster stage of the second maturation division (Fig. 141) we see in
each daughter cell (spermatid) the chromosomes densely apposed, forming together a
rounded, irregular mass, and in only one of the two daughter cells the chromosome x i.
The reduction of the number of chromatin elements in Protenor belfragei is accord-
ingly as follows : Spermatogonium, two univalent chromatin nucleoli, ten univalent
chromosomes, one chromosome x ; first spermatocyte, one bivalent chromatin nucleolus,
five bivalent chromosomes, one chromosome x; second spermatocyte, one univalent
chromatin uucleolus, five univalent chromosomes, one-half chromosome x; spermatid,
one semivalent chromatin nucleolus, five semivalent chromosomes, and either present or
absent one-half chromosome x. This chromosome x is the odd one of the spermato-
gonia ; it does not unite with any other one in the synapsis stage of the spermatocyte,
yet since it sometimes appears bipartite in the synapsis and undergoes a transverse divi-
sion in the first maturation mitosis, it may perhaps be looked upon as bivalent in both
spermatogonium and spermatocyte. If this is a correct conclusion, then the uneven
number of chromosomes in the spermatogonia would be the result of two univalent ones
remaining there united instead of separating this compound, bivalent one being the
chromosome x. This chromosome, as we have seen, behaves in the rest stage of the
spermatogonia like the other chromosomes, but in the growth period of the spermatocytes
it acts in many ways like a chromatin nucleolus.

23. Cymus augiistatus Stal.

Rix testes of this species were studied.

There was no material at my service fixed with Flemming's or Hermann's fluids, so
not being able to use the triple stain of Hermann I was unable to determine the relations
of the chromatin nucleoli in the rest stage of the spermatogonia. The preparations also
showed no favorable cells for counting the chromosomes in this generation.

In the synapsis stage there is a rather small, dumbbell-shaped, and so probably
bivalent, chromatin nucleolus, which becomes spherical in the following (complete) rest
stage of the spermatocyte.

There were no pole views of the chromosomal plate of the first maturation division,
but two pole views of the succeeding dyaster are here given (PL IV, Fig. 143, showing
the chromosomes before taking their definite position in the spindle, while in Fig. 144
they occupy this position and are seen from their ends) ; here can be counted twelve
chromosomes and one smaller body (N. 2, probably the chromatin nucleolus, very small
in Fig. 144). On lateral views of the first maturation monaster (Fig. 142, which, how-
ever, shows only nine of the elements), all the chromosomes usually appear dumbbell-

182 MONTGOMERY A STUDY OF THE CHROMOSOMES

shaped, and so would be bivalent. Sometimes one of them appears rounded or oval
instead of dumbbell-shaped ; this may be a bivalent one seen obliquely, or it might
possibly be a univalent one ; the lack of knowledge of the spermatogonic number does
not permit us to decide which.

24. Ichnodemus falicus Say

Five testes of this species were studied.

Cliromatin nucleoli were not determined in the rest stage of the sperm atogonia.
But they are very probably present there because two small rounded ones can be seen in
the late spermatogonic prophases ; whether more than two I could not determine. All
the testes examined had been fixed with picro-acetic acid, causing such a swelling and
consequent juxtaposition of the chromosomes that in only one case were they sufficiently
separated to be counted (PI. IV, Fig. 145), and here fourteen chromatin elements were
present. Since in the first maturation spindle there are always seven bivalent chromo-
somes, I should think that the fourteen elements of Fig. 145 are univalent chromosomes,
and that in this monaster the chromatin nucleoli are hidden.

In the monaster stage of the first maturation division are seen on pole views (Figs.
147, 148) always seven larger elements, which on lateral view are found to be all dumb-
bell-shaped, and so are probably bivalent. All these seven elements are presumably
chromosomes corresponding to the fourteen found in the spermatogonia (Fig. 145).
Besides these are to be seen at this stage, and also in the preceding prophases (Fig. 146),
two or three smaller elements, which are presumably chromatin nucleoli. Generally
three of these are found (JV. 2, Figs. 146, 147), and generally the case is as in Fig. 146,
two larger and one smaller. The two larger being generally of approximately equal
volume (Figs. 146, 148), it is quite probable that taken together they may represent
one bivalent chromatin nucleolus with separated components. The smaller one some-
times appears transversely constricted, as in Fig. 146, so this one may also be bivalent ;
if this is the case, then there should be four univalent ones in the spermatogonia. But
the number of them could not be determined in the spermatogonia, and in the growth
period of the spermatocytes the stain was not favorable for showing their relations.

25. Peliopelta abbreviate Uhl.

Five testes of this species were studied.

In a pole view of the spermatogonic monaster (PI. IV, Fig. 149, the only clear case
observed) are found sixteen chromatin segments, ten of which are larger and more
elongate than the others, while six are rounded and smaller. The two smallest (N. #)
are probably chromatin nucleoli, by analogy with other species of the Lygceidce.

OF THE GERM CELLS OF METAZOA. 183

In the growth period of the spermatocytes there is one clearly bivalent chromatin
nucleolus, which is frequently apposed to the larger true nucleolus.

In the monaster stage of the first maturation division (Fig. 150, pole view) there are
eight elements, the smallest of which is the chromatin nucleolus (N. 2); on lateral view
all these elements appear dumbbell-shaped and hence are bivalent (Fig. 151). Of the
seven chromosomes of this stage, two are very small and five much larger, the two small
ones in Fig. 151 being designated as a. 2 and b. 2, while in the figure only three of the

large ones are shown. Apparently the two small chromosomes of the spermatocyte cor-
respond to the four small chromosomes of the spermatogonium ; thus the bivalent
chromosomes a. 2 and b. 2 of Fig. 151 would correspond respectively to the univalent
chromosomes a and b of Fig. 149, and the five large bivalent chromosomes of the
spermatocyte to the ten large univalent ones of the spermatogonium. It is very evident
that in the synapsis stage one of the small univalent chromosomes derived from the
spermatogonium never unites with one of the large, for the two univalent components of
each small bivalent chromosome of the spermatocyte have approximately the same vol-
ume. This speaks, of course, very strongly for the maintenance of the individuality of
the chromosomes during these generations.

26. (Edancala dorsalis Say

Four testes of this species were studied.

In the rest stage of the spermatogonium are present two chromatin nucleoli of
rounded form (N. 2, Fig. 152, PL IV) ; these are sometimes attached to one another,
sometimes to the true nucleolus (JV).

In the spermatogonic monaster there are thirteen chromatin elements present, exactly
this number being found in all of nine clear cases. Two of these, which are rounded
and much smaller than the others, are the chromatin nucleoli (N. 3, Figs. 153, 154) ;
the remaining eleven elements are relatively large, elongated chromosomes. All these
elements are halved in the metakinesis.

In the synapsis stage the two chromatin nucleoli combine to form one bivalent one,
and even up to the rest stage of the spermatocyte (Fig. 155) it remains dumbbell-shaped,
with a bridge of linin connecting its univalent components ; it is attached to the surface
cf the true nucleolus (JV, Fig. 155), and the " double nucleolus " so formed usually lies
close to the nuclear membrane.

In the first maturation division there are seven chromatin elements (Fig. 156, pole
view of monaster stage), of which the smallest, usually centrally placed one is the biva-
lent chromatin nucleolus (JV] 2). Of the six chromosomes, five, when seen on lateral
view, are dumbbell-shaped, and so bivalent ; but the sixth is oval in outline without any

184 MONTGOMERY A STUDY OF THE CHROMOSOMES

transverse constriction, about the size of a univalent component of one of the bivalent
chromosomes. This is the chromosome marked x in Fig. 157 ; in this figure is shown
also the chromatin nucleolus (N. 2), but only two of the five bivalent chromosomes.
This sixth small chromosome is univalent and unipartite, and evidently is the odd,
eleventh, chromosome of the spermatogonium, which had no fellow to combine with
during the synapsis. It is always recognizable on lateral views (Fig. 157) by its peculiar
volume and form, and even on pole views of the chromosomal plane is recognizable by
its lesser depth (Fig. 156, x).

The first maturation division halves all seven elements (Fig. 158, anaphase), being
a transverse (reducing) division of the five bivalent chromosomes and of the chromatin
nucleolus, but in what plane the univalent chromosome (x) divides could not be deter-
mined on account of its nearly spherical form. Apparently also in the second matura-
tion division all the six chromosomes become divided, since in the spermatid six chroma-
tin elements can frequently be counted ; but I am not certain that the sixth chromosome
does become divided in this mitosis.

The reduction in the number of chromosomes for this species is accordingly :
/Spermatogonium, two univalent chromatin nucleoli, eleven univalent chromosomes ; first
spermatocyte, one bivalent chromatin nucleolus, five bivalent chromosomes, one univalent
chromosome ; second spermatocyte, one univalent chromatin nucleolus, five univalent
chromosomes, one semivalent chromosome.

27. Oncopeltus fasciatus Dall.

Eight testes of this species were studied.

The rest stage of the spermatogonium (PL IV, Fig. 159) shows usually one com-
paratively large, elongate chromatin nucleolus (N. 2), which is generally peripheral in
position ; this apparently represents two joined end to end, for sometimes two separate
ones can be seen.

In the spermatogonic monaster there are sixteen chromatin elements (Fig. 160).
Fourteen are chromosomes and two are chromatin nucleoli, as the relations in the sperma-
tocyte mitoses will demonstrate. But it is difficult to determine which two are the
chromatin nucleoli, all sixteen elements being of approximately equal size, though,
judging by analogy with the other species of the family, they are probably the smallest
two (N. 2, Fig. 160).

In the synapsis the fourteen chromosomes unite to form seven bivalent ones. But
there is never any very close union of the two chromatin nucleoli, and in the rest stage
of the spermatocytes the following conditions are found : (1) the chromatin nucleoli
apposed to one another and to the true nucleolus (Fig. 161) ; (2) apposed to one

OF THE GERM CELLS OF METAZOA. 185

another, but separated from the nucleolus (Fig. 162) ; (3) the chromatin nucleoli sepa-
rate from one another, and only one in contact with the nucleolus (Figs. 163, 165) ; (4)
the chromatin nucleoli separate from one another and from the nucleolus ; (5) the
chromatin nucleoli separate from one another, but both attached to the nucleolus (Fig.
164). These conditions do not appear to be stages in position, but rather individual
variations. Whenever the two chromatin nucleoli are mutually apposed, it is never an
intimate apposition i. e., they never fuse to form one large rounded one such as is the
general rule in the Pentatomidce. Through the rest stage of the spermatocytes each
chromatin nucleolus remains elongate (they appear round only when seen from the end).
Sometimes each shows a trace of a transverse constriction, and in one case (Fig. 165) one
of them was separated into two parts.

Through the prophases of the first maturation mitosis the chromatin nucleoli remain
separate from one another, and each is elongate in form (Fig. 166, N. 2, showing also all
seven bivalent chromosomes on lateral view).

Pole views of the monaster stage of the first maturation division show in most cases
nine chromatin elements (Fig. 167) ; the seven larger ones are chromosomes, all bivalent
and on lateral view dumbbell-shaped (Fig. 166); the two smaller, centrally placed ones
are the two chromatin nucleoli (N. 2, Fig. 167). In one case, ten elements were seen on
pole view (Fig. 168) ; this is very unusual, but it may probably be explained by the
assumption that there are here two chromatin nucleoli, six bivalent chromosomes, and
two univalent chromosomes formed by the precocious separation of the parts of a biva-
lent chromosome. A lateral view of the spindle is given (Fig. 169), showing the two
chromatin nucleoli, but only two of the seven bivalent chromosomes.

In the metaphase of this mitosis all seven chromosomes become transversely divided
(reduction division), and each of the chromatin nucleoli becomes also divided in a plane
perpendicular to its long axis (compare Figs. 169 and 170, in each of which only two of
the seven chromosomes are shown). A pole view of one cell of the dyaster stage followr
ing (Fig. 171) shows the two daughter chromatin nucleoli (N. 2) and the seven daughter
(univalent) chromosomes (the apparent transverse constrictions on them representing the
reappearance of the original longitudinal split ; compare the left-hand chromosome of
Fig. 170).

The behavior of the chromatin nucleoli is thus different from that of the other
Hemiptera examined in the following regards : (1) their large size in the spermatogonic
monaster (Fig. 160), so that they can hardly be distinguished in volume from the
chromosomes ; (2) the phenomenon that they remain more or less separate from one
another during the growth period of the spermatocytes (Figs. 161-165) and prophases
of the first maturation division (Fig. 166) ; (3) the fact that one or both may appear

180 MONTGOMERY A STUDY OF THE CHROMOSOMES

bipartite in the growth period (Fig. 165) ; and (4) the fact that they remain separate
from one another in the first maturation division, and that each divides transversely
(Figs. 167, 169, 170, N. #). The last mentioned point deserves particular consideration,
for a transverse division of a chromatin element in the Hemipto-a always means a reduc-
tion division /. e., a separation of two whole nnivalent components of one already biva-
lent element. From these facts we are led to the conclusion that here the chromatin nu-
cleoli are virtually bivalent in the spermatogonia, and that since the spermatogonic division
gives a longitudinal half of each of them to the spermatocytes, that each is already
bivalent in the spermatocytes & bivalence then produced before the synapsis stage of
the growth period. This conclusion would explain all their peculiarities listed above.
In the spermatogonium accordingly there would be virtually four chromatin nucleoli,
twice the number found in the other species of the Lygci'ldce (with possibly the exception
of the not fully explained Ichnodemus falicus).

CAPSIDJE.

/

28. Leptopterna dolabrata Linn.

Three testes of this species were studied.

There were no spermatogonia on any preparations (taken from adults in the last in-
star before copulation).

In the rest stage of the growth period of the spermatocytes are found the following
relations for the chromatin nucleoli : There are two chromatin nucleoli, which (1) are
attached to one another but separate from the true nucleolus (Plate IV, Fig. 172) ; (2)
they are attached together and to the true nucleolus (Fig. 173) ; (3) they are separated
from one another but both attached to the true nucleolus (Fig. 175), often at opposite
poles of the latter (Fig. 174). In almost all cases they are attached to the true nucleolus, so
that Fig. 172 represents an unusual case. Each chromatin nucleolus is probably a nniva-
lent one, for it never shows a bipartite appearance and is usually rounded, so that in Figs.
172 and 173 the two together would constitute one bivalent chromatin one ; but there
cannot be certainty on this point until the number in the spermatogonia is determined.

Pole views of the monaster stage of the first maturation division (Fig. 176) show
seventeen chromatin elements, one of which, centrally placed, is always much larger than
the others. On lateral view all appear dumbbell-shaped and so are probably biva-
lent. Probably one of these elements represents the bivalent chromatin nucleolus de-
scribed for the growth period, then the sixteen remaining would be chromosomes.

OF THE GERM CELLS OF METAZOA. 187

29. Calocoris rapidus Say

Three testes of this species were studied.

The number of chromatin nucleoli in the rest stage of the sperm atogonia was not
determined.

In the most favorable pole view of a spermatogonic monaster were counted about
thirty chromatiu elements (Plate IV, Fig. 177), but these elements were densely grouped
so that I could not be positive as to the exact number. Since there are in the sper-
matocytes fourteen bivalent chromosomes, two bivalent chromatin nucleoli and one that is
probably univalent, there would be probably in the spermatogonia twenty-eight univalent
chromosomes and five univalent chromatin nucleoli, a total of thirty-three elements.

In the telophase and rest stages of the spermatocytes there is a large true nucleolus,
which is remarkable in being flattened against the nuclear membrane (N, Figs. 178, 179);
it appeai-s sickle-shaped on cross section, and has irregularly lobular outlines on surface
view. In these stages there are five small chromatin nucleoli (N. 2, Fig. 178) ; one of
these is larger than the others and always spherical in form (the larger one of Figs. 178
and 179), and since it never appears bipartite is presumably univalent; it is frequently
attached to the true nucleolus. The four other chromatin nucleoli are arranged in two
pairs, the two components of a pair being connected by a band of linin (Figs. 178, 180,
only one of the pairs shown in Fig. 179). Each one of. these four is very small and
spherical, and accordingly probably univalent, and each pair would then be bivalent.
Thus there would appear to be in the spermatocytes one larger univalent chromatin
nucleolns and two bivalent ones, in each of the latter the univalent components being not
closely apposed.

Pole views of the monaster stage of the first maturation division show always sixteen
chromatin elements (Figs. 185, 186). Three of these are always distinguishable by their
much smaller size (N. 2, Figs. 185, 186). These three probably represent the three
chromatin nucleoli of the preceding growth period ; two of them appear dumbbell-shaped
on lateral view (Fig. 181, and one of the two is shown in Fig. 184), obviously representing
the two bivalent ones of the growth period ; while the third one always appears rounded
and never dumbbell-shaped (Fig. 181, the lowest of the elements designated N. <?), and on
lateral view of the spindle lies nearer one pole of the spindle than the other (N. ,2 Fig.
183), this one obviously representing the univalent chromatin nucleolns of the growth
period. Figs. 181-184 represent oblique lateral views of the spindles, so that in each case
only one spindle pole is shown, such oblique views giving the best views of the chromatiu
elements. The thirteen remaining larger elements of the first spermatocytes are chromo-
somes, and lateral views show that twelve of these are dumbbell-shaped and hence prob-
ably bivalent ; but the thirteenth is quadrivalent, composed of two bivalent (dumbbell-

188 MONTGOMERY A STUDY OF THE CHROMOSOMES

shaped) placed side by side with their long axes parallel. This quadrivalent chromosome
shows its nature plainly on lateral view (t, Figs. 182, 183) ; on pole view it may
always be told by its greater volume (t, Figs. 185, 18G), sometimes even on pole view it
appears slightly constricted (Fig. 186, /), the constriction then denoting the plane of appo-
sition of the two bivalent chromosomes of which it is composed; it is placed in the spindle
so that the tranverse constriction of each of its component chromosomes lies in the piano
of the equator (Figs. 182, 183).

There are accordingly in the first spermatocyte two bivalent chromatin nucleoli, one
univalent chromatin nucleolus, twelve bivalent chromosomes, and one quadrivalent
chromosome, in all sixteen chromatin elements.

In the spindle of the second spermatocyte are found either fifteen chromatin elements
(Fig 188) or sixteen (Fig. 187). This disparity in number is produced by the unival-
ent chromatin nucleolus not dividing in the first maturation mitosis but passing undivided
into one of the two daughter cells (second spermatocytes), for it will be remembered that
in the monaster stage of the first mitosis it always lies a little outside of the plane of the
equator, nearer one pole of the spindle than the other (Fig. 183, N. 2). The rest of the
chromatin elements of the second spermatocyte are univalent halves of those in the first
spermatocyte (the first maturation division is a reduction division), namely, halves of the
two bivalent chromatin nucleoli, of the twelve bivalent chromosomes, and of the one
quadrivalent chromosome. The latter element can be recognized in the second spermato-
cytes by its greater size (/, Figs. 187, 188), and it here consists of two univalent chromo-
somes placed in apposition ; in the spindle of the first maturation mitosis (Figs. 182, 183)
it was so placed that each of its bivalent chromosomes underwent a transverse (reduction)
division, just as was the case with the other bivalent chromosomes.

30. Pcecilocapsus lineatus Fabr.

Six testes of this species were studied.

There were no spermatogonic mitoses on my preparations, and I could not determine
the relations of the chromatin nucleoli in the rest stage of the spermatogonia.

In the rest stage of the spermatocytes are found two chromatin nucleoli (N. 2, Fig.
189, PI. IV) ; in one testis two additional, very minute chromatin uucleoli seemed
to be present, but I could not find them on the other preparations. One of the chromatin
nucleoli is very large and clearly bilobed (Fig. 189) so that it would seem to be bivalent ;
the other is considerably smaller and apparently always rounded, so that it may be uni-
valent. These two chromatin nucleoli are sometimes, but not usually, in mutual contact.

Pole views of the monaster stage of the first maturation mitosis show always eighteen
chromatin elements (Fig. 190). One of these is much smaller and one much larger than

OF THE GERM CELLS OF METAZOA. 189

the others. Probably two of the eighteen elements correspond to the two chromatin
nucleoli of the growth period, but I cannot determine which two they are ; if this is so,
then there would be here sixteen chromosomes (all apparently bivalent judging from their
dumbbell-shape on lateral view), and one bivalent and one univalent chromatin nucleolus.
But there can be no surety in regard to these valences without a knowledge of the num-
bers in the spermatogonia.

31. Pcecilocapsus goniphorus Say

Four testes of this species were studied.

There were no spermatogonic mitoses on my preparations (all from adult individuals).

In the rest stage of the spermatocytes there are present four bipartite chromatin
nucleoli (N. 2, Figs. 192-195, PI. V), in one single case there were five (Fig. 191). The
largest of the four is composed generally of two rods placed side by side (this is shown
in lateral view in Figs. 191, 194, in end view in Figs. 192, 193, 195) ; the lateral view of
this one (the lower of the large ones of Fig. 191) sometimes shows that each of its com-
ponent rods may be transversely constricted, which might imply that each of the rods
is bivalent and hence that the whole is quadrivalent. In each of the three small
bipartite chromatin nucleoli the univalent components may be closely apposed to one an-
other (Figs. 191, 192), or may be more or less widely separated (Figs. 193-195). Of the
four chromatin nucleoli the largest and the smallest are generally attached to opposite
poles of the true nucleolus (N. 2, Figs. 191-194), though the relative positions vary con-
siderably as shown by the figures ; and it is the two which are generally not so attached
which have their component parts most widely separated.

Thus there are at least four, possibly five, bivalent chromatin nucleoli in the sperma-
tocytes.

Pole views of the monaster stage of the first maturation mitosis show either seventeen
chromatin elements of approximately equal volume (Fig. 197), and this was the rule in
two of the testes examined ; while in a third testis in the majority of cases there was
present a smaller element in addition to the seventeen larger ones (t, Fig. 196); possibly
this small element may always be present, but frequently escape observation by being
closely apposed to one of the larger elements. All these elements appear dumbbell-shaped
on lateral view (Fig. 198), and so are probably bivalent.

How may of these eighteen elements are chromosomes and how many are chromatin
nucleoli I cannot determine, since the number in the spermatogonia was not ascertained.
Possibly the small element (t of Fig. 196) may represent the largest chromatin nucleolus
of the preceding growth period, and the seventeen remaining be chromosomes ; or if all
four chromatin nucleoli are represented in the first maturation division, then there would
remain fourteen chromosomes.

190 MONTGOMERY A STl DY OF THK CHROMOSOMES

PHYMAT1D.K.

32. Phymata sp. (P. wolffii Stal. ?).

Nine testes of this species were studied.

In the rest stage of the sperraatogonium there are two chroraatin nucleoli (JV. .',
Fig. 199, PI. V), usually unequal in size ; frequently one or both of them is in contact
with the true nucleolus (jV).

In the spermatogonic monaster are seen on pole view (Fig. 200) thirty chromatin
elements; two of these certainly represent the chromatin nucleoli, but they offer no pecu-
liarities by which they can be distinguished from the twenty-eight small chromosomes.

In the synapsis stage the twenty-eight chromosomes unite to form fourteen bivalent
ones, and the two chromatiu nucleoli to form one bivalent one. The latter is in the telo-
phase and rest stage of the spermatocyte (Fig. 201) clearly transversely constricted, show-
ing its bipartite nature, and is always apposed to the surface of the larger true nucleolus
(Fig. 201, JV.).

Pole views of the monaster stage of the first maturation division (Fig. 202) always
show exactly fifteen chromatin elements, namely, fourteen chromosomes and one chroma-
tin nucleolus. All these elements are found to be dumbbell-shaped on lateral view (Fig.
203, showing eight of them), so that all are bivalent. The chromatin nucleolus cannot be
distinguished in size from the chromosomes.

NABID.E.
33. Coriscus ferus Linn.

Three testes of this species were studied.

The relations of the chromatin nucleoli were not determined in the small nuclei of
the spermatogonia, and the spermatogonic mitoses were not favorable for counting the
small, rounded chromosomes.

In the growth period of the spermatocytes there is usually one large bivalent
chromatin nucleolus, with its component parts in close apposition (N. 2, Fig. 205, PI. V),
attached to the true nucleolus (JV). Sometimes its component parts (which are of equal
volume) are separated from one another, and then both may be attached to the same true
nucleolus (Fig. 204), or they may be apposed to separate nucleoli, or only one of them
may be apposed to a nucleolus (of which there are generally two, sometimes three). Be-
sides the large, bivalent chromatin nucleolus can be seen in most nuclei a much smaller
chromatin nucleolus (Fig. 205) which stains like the larger one ; it is generally close to
the nuclear membrane, but is occasionally apposed to a true nucleolus.

In the first maturation division there are ten chromatin elements (Fig. 206, in which

OF THE GERM CELLS OF METAZOA. 191

five are seen on lateral view and five on pole view, all these elements not having taken
their definite position in the equator of the spindle). All are bivalent, as is proved by
their bipartite appearance on lateral view. One of them is smaller than the others, and
may represent the large (bivalent) chromatin nucleolus of the growth period (N. 2, Fig.
206), while the nine remaining elements are probably chromosomes ; if this interpretation
be correct then the small (imivalent ?) chromatin nucleolus found in the growth period
would not be represented in the first maturation division.

REDUVIID.E.
34. Acholla multispinosa De G.

Eight testes of this species were studied.

The relations of the minute chromatin nucleoli could not be determined in the rest
stage of the spermatogoni;i.

In the most favorable pole view of a spermatogonic monaster (PI. V, Fig. 207) could
be counted thirty-one chromatin elements. The twenty-four larger elements seen here are
univalent chromosomes, the seven smallest are chromatin uucleoli (some of them very
minute). Of the latter there are six arranged in three pairs and one that is isolated.
Now we shall find that in the spermatocytes there are four bivalent chromatin nucleoli
(Figs. 208-211), so that in the spermatogonia there should be eight ; and accordingly
though only seven are to be seen in Fig. 207, we are justified in concluding that an eighth
must be present there but hidden from view by one of the chromosomes. Thus there
would be in the spermatogonium in all probability twenty-four chromosomes and eight
chromatin nucleoli, in all thirty-two elements.

In the syuapsis stage of the growth period the twenty-four chromosomes unite to
form twelve bivalent ones. The eight chromatin nucleoli likewise combine to form four
bivalent ones, which near the close of the growth period (Figs. 208, 209) are seen to be
small bodies attached to the surface of the true nucleolus (N). Each is dumbbell-shaped,
usually with its component univalent parts in close apposition, but occasionally the latter
are more or less separated from one another.

Pole views of the monaster stage of the first maturation division (Fig. 210) show six-
teen elements, namely, twelve larger, bivalent chromosomes (dumbbell-shaped on lateral
view) and four much smaller chromatin nucleoli (N. 2}. All these elements are halved
in the following metakinesis, which is a transverse (reduction) division, and pole views of
the daughter cells (second spermatocytes, Fig. 211) show twelve larger, univalent chromo-
somes and four smaller, univalent chromatin nucleoli (N. 2}.

192 MONTGOMERY A STUDY OF THE CHROMOSOMES

35. /Sinea diadem a Fabr.

t

Two testes of this species were studied.

The relations of the chromatin nucleoli iu the rest stage of the spermatogonia could
not be determined.

There were on ray preparations only a few spermatogonic monaster stages, and none
of these were favorable for determining the number of chromosomes ; the chromosomes
are small, densely grouped, and particularly minute elements among them might be
chromatin uucleoli.

In the rest stage of the spermatocyte are present four small chromatin nucleoli (N.
2, Figs. 212, 213, PI. V), all of them attached to the surface of a large true nucleolus (N}.
One is larger than the others, and appearing on lateral view to be always transversely con-
stricted (Fig. 213) is probably bivalent ; and as one of the three smaller ones is some-
times found to be bipartite, this too would be bivalent. The two remaining are appar-
ently always spherical and not transversely constricted, so that they would seem to be
uuivalent. If this interpretation is correct, there would be here two bivalent chromatin
nucleoli each with its component parts closely apposed, and one bivalent one with its
components separated, that is three bivalent ones in all ; and in the first maturation divi-
sion there are three chromatin nucleoli present (Figs. 214, 215), which would corroborate
this conclusion. But since it could not be determined how many there are in the sperma-
togonia, the relations in the spermatocytes cannot be considered positively demonstrated.

Pole views of the monaster stage of the first maturation division (Figs. 214, 215)
show always three minute chromatin elements, which are chromatin nucleoli (N. 2], and
apparently thirteen larger chromosomes. But a careful examination of these larger
elements shows that in every case four bivalent chromosomes form together a plurivalerit
one, all four being closely apposed and with their long axes parallel to one another.
This is best seen on lateral view of the spindle (Figs. 217, 218), where the four chromo-
somes marked t are always found to be grouped close together ; in Fig. 216 only these
four chromosomes with their mantle fibre attachments are seen on lateral view. Of the
four chromosomes thus grouped together, the two middle ones stand in the closest apposi-
tion (Fig. 210-218) ; and on pole views these two middle ones may be so closely apposed
as to appear as one long one (t, Fig. 214), or a slight transverse constriction marks the
division line between them (/, Fig. 215). It follows accordingly that the three elements
seen on pole view and marked t are really four bivalent chromosomes, the two central ones
being so closely apposed as to appear generally as one long one. A comparison of the
elements designated t in Figs. 214 and 215 with the corresponding elements in Figs. 216-
218 makes this evident. Accordingly there are really fourteen bivalent (dumbbell-shaped)

OF THE GERM CELLS OF METAZOA. 193

chromosomes present, four of which are always grouped together to form a particular
group.

The metakinesis halves each of the fourteen bivalent chromosomes transversely
(reduction division, Figs. 217, 218), but the four which together form the group that has
been described divide later than the others, as can be seen from the figures given ; and it
is at this metaphase that this group of four can be most easily recognized. Such a group
of four closely united bivalent chromosomes has not been found by me in any other
Hemipteron.

36. Prionidus cristatus Linn.

Four testes of this species were studied.

In the rest stage of the sperm atogonium there are apparently five chrornatin uucleoli
(N. 2, Figs. 220-222, PI. V), two of which are generally considerably larger than the
others, and some or all of which may be attached to the true nucleolus (N, Figs. 221,
222). But it is difficult to be sure of the exact number, for sometimes not more than
four can be seen (Fig. 221).

In the spermatogonic monaster on pole views (Figs. 223, 224) are seen twenty-six
chromatin elements, three of which (N. 2} are always much smaller and (by analogy
with the other species of the Reduviidce) would probably represent the three small
chromatin nucleoli of the preceding rest stage. Two of the twenty-six chromatin ele-
ments are elongate in form and much larger than the others ; these may be chromosomes
(which would be more probable), or they might represent the large chromatin nucleoli of
the rest stage of the spermatogonium. If there are really five chromatin nucleoli in the
spermatogonia, then there would be these five elements and twenty-one chromosomes
present in the spermatogonic mitosis. In five pole views I could count exactly twenty-
six chromatin elements, in two there were either twenty-six or twenty-seven, and in one
either twenty-four or twenty-five ; but in all these cases three particularly small and
two particularly large elements could be distinguished.

In the rest stage of the spermatocyte (Fig. 225) are found four chromatin uucleoli
{N. 2} attached to the true nucleolus (N). One of these is longer than the others and
rod-shaped, and may represent the two larger chromatin nucleoli of the spermatogonium
joined into one bivalent one ; the three smaller ones may appear rounded or slightly
elongate, and these may represent the three small chromatin nucleoli of the spermato-
gonium ; occasionally two of the small ones are apposed together.

There were no maturation mitoses on my preparations (from adults taken in the
month of September).

194 MONTGOMERY A STUDY OF THE CHROMOSOMES

37. Milyas cincius Fabr.

In the single testis of this species studied (an individual of the mouth of September)
there were no mitoses.

In the rest stage of the spermatocytes (Figs. 226-228) are found one long, rod-
shaped chromatin nucleolus and two or three smaller ones (N. 2), all apposed to the true
nucleolus (N). The long one is certainly bivalent, since it frequently shows a transverse
constriction (Fig. 228) or is bent at the middle point (Fig. 227); on account of the
length of each of its component parts it might be concluded that each of them is biva-
lent i. e., that the whole element is quadrivalent but neither of the parts appear bipar-
tite, so that this long chromatin nucleolus would more probably be bivalent. In those
cases where only two smaller chromatin nucleoli are present (Fig. 228), each of them is
clearly transversely constricted (bipartite) and hence bivalent ; where three are present
we find that one is bipartite and accordingly bivalent (Fig. 227), the other two are
spherical and hence univalent. By comparing Figs. 227 and 228 we find that the two
spherical chromatiii nucleoli of the former would together represent one of the small
bivalent chromatiii nucleoli of the latter. Thus we may conclude that there are three
bivalent chromatin nucleoli present in these spermatocytes, though the two components
of one of them may be separated.

LIMNOBATIDJS.

38. Limnobates lineata Say.

Of this small species I was able to procure the testes of only one (adult) individual.
The only stages of sperrnatogenesis present were spermatocytes in the growth period.

In the large spermatocytes in the rest stage (PL V, Fig. 219) is found a large
chromatiu nucleolus (N. 2), usually apposed to the true nucleolus (N) in such a manner
that the chromatin nucleolus touches with one pole the nuclear membrane, with the other
the true nucleolus.

HYDROBATID.E.

39. Hygoirechus sp.

Twelve testes were studied of this species (from the vicinity of Philadelphia).

The relations of the chromatin nucleoli in the rest stage of the spermatogonia could
not be determined, since the nuclei are very small at this stage.

The spermatogonic monaster shows on pole view (PI. V, Fig. 229) exactly twenty
chromatin elements, of which the eighteen largest are chromosomes and the two smallest
(N. 2) probably chromatin uucleoli.

In the synapsis stage of the spermatocyte (Fig. 230) are seen two small chromatin

OF THE GEKM CELLS OF METAZOA. 195

nucleoli (N. 3), which are not closely apposed ; the true nucleolus (N) is much larger.
In the rest stage the chromatiu nucleoli are usually widely separated from one another,
and on account of their small size are difficult to discover. In the growth period they
do not stain bright red with the saffranine-gentian violet stain of Hermann, but deep
violet, even on excellently stained preparations. Such a staining reaction as this I have
not found for the chroma-tin nucleoli of other ffemiptera, for in all the other forms
examined the chromatin nucleoli take the red saffranine stain intensely even while the
chromosomes have taken the violet stain (in the rest stage). Perhaps in Hyc/otrechus
the chromatin nucleoli differ chemically less from the chromosomes than in the other
ffemiptera, and undergo in the growth period changes parallel to those of the chromo-
somes.

Pole views of the monaster stage of the first maturation division (Fig. 231) show
always exactly eleven chromatin elements. On account of the number in the spermato-
gonia (twenty), I would interpret these as nine bivalent chromosomes, and two uuivalent
chromatin nucleoli which are not combined into one bivalent one. This is very probable,
since the two univaleut chromatin nucleoli are often widely separated in the growth
period ; and on lateral views of the monaster stage of the first maturation division, there
are found two bodies which are spherical and not dumbbell-shaped. On no lateral view
of this monaster stage could I see all the nine chromosomes clearly ; but in one case I saw
eight of them, all clearly dumbbell-shaped (bipartite), so that probably all nine are
bivalent. The first maturation division is reductional, as in the other Hemiptcm.

40. Limnotrechiis marginatus Say

Two testes of this species were studied.

There were no spermatogonic monaster stages present, and I could not determine the
relations of the chromatiu nucleoli in the rest stage of these cells.

In the rest stage of the spermatocyte (PI. V, Fig. 232) is found a large true nucleo-
lus (N), and separated from it, usually close to the nuclear membrane, a rounded
chromatin nucleolus (N. 2).

Pole views of the first maturation monaster (Fig. 233) show eleven chromatin ele-
ments ; sometimes one appears much smaller than the others, and it may represent the
chromatin nucleolus.

NAUCORID.E.

41. Pelocoris femorata Pal. Beauv.

Fourteen testes of this species were studied from adult and half-grown individuals
of different seasons of the year. There were an abundance of rest stages of spermato-
gonia and of spermatocytes in the growth period, but no maturation mitoses present.

196 MONTGOMERY A STUDY OF THE CHROMOSOMES

The chromosomes ^ere counted in a number of pole views of the monaster stage of
the spermatogonia, but in most cases they were so densely grouped as to make the
numbers obtained very uncertain. In the most favorable case (PI. V, Fig. 234) appa-
rently twenty chromatin elements are present, but I could not be certain of this number.

The relations of the chromatin nucleoli in the growth period are very puzzling.
On preparations stained with Hermann's saffranine-gentian violet, there appear to be a
variable number of rounded bodies of different volume which stain bright red ; some-
times they are arranged in pairs, sometimes in long chains, sometimes they show no
regular arrangement whatsoever. One nuclear body, much larger than the others and
generally irregular in outline, may be a true nucleolus. If all the red-staining bodies
are chromatin nucleoli, they would seem to be present in an unusually large number in
this species. A study of further material will be necessary to explain the nature and
relations of these bodies.

BELOSTOMATID.E.

42. Zaitha sp.

There are two species of this genus known in the vicinity of Philadelphia where I
collected my material, namely, Z. fluminea Say and Z. aurantiacum (Leidy), but which
species it was that I collected I omitted at the time to determine. Ten testes were
examined.*

In the rest stage of the spermatogonium (Fig. 235) are present two small chromatin
nucleoli (N. #) apposed to a large true nucleolus (N).

Pole views of the spermatogonic monaster stage (Figs. 236, 237) show twenty-four
chromatin elements, of which the smallest two (N. 3} represent the chromatin nucleoli,
and the remaining twenty-two are chromosomes. Of the chromosomes, four are always
elongate and much larger than the others (t, Figs. 236, 237).

In the synapsis stage of the growth period the twenty-two chromosomes unite to
form eleven bivalent ones. In the rest stage of the spermatocytes there are two uuiva-
lent chromatin nucleoli, sometimes joined to make a bivalent one, attached to the surface
of the true nucleolus.

Pole views of the monaster stage of the first maturation division (Fig. 238) show

* In my " Note on the Genital Organs of Zaitha " (American Naturalist, Vol. xxxiv, 1900). I described the
structure of these testes. To the text figure B given iu that paper, I would add now that the earlier stages of
spermatogenesis are to be found in what I called the " terminal fibres " at the proximal end of the testis, these fibres
being five long and much convoluted slender tubes which interlace together and form a rounded whitish mass at
the extreme proximal end of the testis ; they were not correctly represented iu the figure cited. In adult individuals
it is only in this portion of the testis that the earlier spermatogenetic stages occur, all the rest of the testis being
filled with spermatozoa. It is necessary to collect individuals in the month of May (shortly before the copulation),
in order to obtain the stages of the maturation divisions.

OF THE GERM CELLS OF METAZOA. 197

always thirteen cliromatin elements. Two of these are much smaller than the others
(W. 2), and are obviously the univaJent cliromatin nucleoli, which at this stage do not
make up a bivalent one. The eleven large elements are Chromosomes, and since all of
them appear dumbbell-shaped on lateral view, they are all bivalent. Two of the eleven
chromosomes have a markedly greater volume than the others (t 3, Fig. 238), and
these correspond to the four large chromosomes of the spermatogonia (t, Figs. 23G, 237) ;
that is to say, in the syuapsis the four large chromosomes derived from the spermatogo-
nia unite together to form two bivalent ones, and a large one never appears to unite
with a small one.

The first maturation mitosis is a reduction division, and each daughter cell (second
spermatocyte) receives eleven whole uuivalent chromosomes.

III. GENERAL CONCLUSIONS.

1. The process of spermatogenesis in the Hemiptera, and the individuality of the
chromosomes.

In the Hemiptera heteroptera we find, as generally elsewhere in the Metazoa, a
number of generations of spermatogonia in each of which all the cliromatin elements
are halved in metakinesis (apparently in all cases equationally), the last generation pro-
ducing spermatocytes of the first order. These spermatocytes enter upon a growth
period of long duration, which is followed by the first maturation division, resulting
in the formation of spermatocytes of the second order ; and in the second maturation
division the spermatocytes of the second order are divided into spermatids. There are
always exactly two maturation divisions and no more. The metamorphosis of the
spermatids into the spermatozoa has not been studied by me, and has not the broad com-
parative interest of the preceding stages, but it has been described by Henking (1890)
and Paulmier (1899).

The growth period of the spermatocytes is of the greatest interest, for here are a
remarkable series of changes not found in any other generation of the germ cells nor, so
far as is known, in any somatic cells. And the most important of these changes are
found in the syuapsis stage of the growth period. The synapsis stage is well-marked in
all the Hemiptera examined by me without exception, characterized by a dense grouping
of the cliromatin loops ; the citations given by me in my study on Peripatus (1901)
show that it seems to be present in almost all, if not all the Metazoa in which the
spermatogenesis and ovogenesis has been carefully examined. The dense grouping of
the chromosomes in this stage is not an artifact produced by faulty fixation methods, as
McC'lung (1900) has recently maintained, for exactly the same appearances are to be

198 MONTGOMERY A STUDY OF THE CHROMOSOMES

found after the action of most diverse fixatives. The dense and interlacing grouping of
the chromosomes that is so characteristic for the sy-napsis of Insects, Gopepods, Ascaris
and some other forms, is, however, not found in all ; thus it does not appear to occur in
Salamandm (Meves, 189G).

Moore (1895) first gave the name " synaptic phase " to that stage in the growth
period of Elasmobranchs when the reduction in the number of the chromosomes takes
place. Accordingly, the criterion of the synapsis stage is first of all the combination of
univalent chromosomes to form bivalent ones ; whether the chromosomes are then densely
grouped or not is of secondary importance. A special chapter of the present paper is
given to the broader significance of this stage.

In all the Hcmiptera examined, and also in Peripatus, I have found that bivalent
chromosomes are formed in the synapsis by a union, end to end, of every two univalent
chromosomes. All other writers on this stage, with the exception of Brauer, have been
unable to determine how the univalent chromosomes become united together; Brauer's
(1893b) careful study of the growth period in the spermatogenesis of Ascaris rendered
it very probable that it is a union end to end, but his figures do not prove it absolutely.
Certain writers state that the reduction of the number in chromosomes is effected by the
chromatin spireni segmenting into only half the normal number of chromosomes. This
is, however, an incorrect statement, inasmuch as the reduction in number is occasioned in
some cases (Hemvptera, Penpaius) before the " spirem " stage of the first maturation
division, and inasmuch as in most cases, if not all, there is no continuous chromatin
spirem found at any time during the growth period and prophases of the first maturation
mitosis.

Accordingly in the Hemiptera the reduction in number of the chromosomes is
effected in the synapsis stage, a long while before the maturation divisions, by a union
end to end of every two chromosomes. During the synapsis stage the chromosomes
become split longitudinally, as was first shown by Paulmier (1898, 1899) for Anasa
a process that I had overlooked in my former paper (1898). Each bivalent chromosome
is thus both transversely and longitudinally split before the maturation mitoses, the
transverse split represented by the band of linin joining the approximated ends of the
two univalent chromosomes.

At the close of the growth period there is a well-defined rest stage in most Ifc/>tijt/<'/'<i,
when chromosomal boundaries are practically indistinguishable ; but in the Coreidve and
Reduviidce there appears to be no such stage, and accordingly such a stage would appear
to have no broad significance.

In the early prophases of the maturation divisions the chromosomes are bivalent
but quadripartite, each one being transversely and longitudinally split ; in the later pro-

OF THE GERM CELLS OF METAZOA. 199

phases, up to the monaster stage of the first mitosis, the longitudinal split generally closes
temporarily. The definitive form of these chromosomes in all the Hemiptera examined
is that of a dumbbell which may be either straight or bent ; ring forms are more infre-
quent, but are occasionally found in all species. In the Hemiptera, as in Peripatus,
each ring may be conceived as a dumbbell which has become bent until its ends meet,
and accordingly the hollow of the ring is not the longitudinal split, but a space sepa-
rating the univalent chromosomes. That is to say, generally only one end of one
univalent chromosome is joined to one end of the other, but in the ring form both ends
of the one are joined with both ends of the other. In all bivalent chromosomes that
have not the ring form a longitudinal axis can be plainly determined, and generally each
univalent chromosome is elongated in the same line ; the constriction perpendicular to
this long axis is a true transverse division, and is the band of linin joining the ends of
the two univalent chromosomes. This orientation of the axes of the bivalent chromo-
somes allows the positive determination of the manner in which the chromosomes are
halved in the maturation divisions.

In the first maturation division in all the Hemiptera examined the bivalent chromo-
somes are transversely divided, i. e., whole univalent chromosomes are separated ; in the
anaphase of this division the longitudinal split of the univalent chromosomes reappears,
and in the second maturation mitosis the univalent chromosomes are halved through the
plane of this split (equation division). The valences of the chromosomes in the successive
generations are accordingly : spermatogonium, univalent ; first spermatocyte, bivalent ;
second spermatocyte, univalent ; spermatid, semivalent. The classing of the chromosomes
as semivalent in the spermatid may appear surprising, for they have always been consid-
ered univalent ; but they must be considered semivalent with reference to the number in
the spermatogonia in those cases, as in all the Hemiptera examined, where the second
maturation division follows immediately upon the first without any indication of an inter-
mediate rest stage.

Thus in the Hemiptera, as in Peripatus, the maturation divisions do not accomplish
the reduction in number of the chromosomes, for this takes place long before in the
growth period; the first maturation division separates entire univalent chromosomes
(pseudoreduction, Riickert, 1894), the second halves each univalent chromosome equation-
ally, and thus halves the chromatin mass. Though the chromosomes of the spermatid are
logically semivalent with reference to those in the spermatogonia, yet they are potentially
univalent on account of the increase in mass of the chromatin during the growth period
(where at least a doubling of the mass occurs).

In my study on Peripatus (1901) it was shown that the individuality of the chromo-
somes is maintained from the last sperm atogonic mitosis up to and through the maturation

200 MONTGOMERY A STUDY OF THE CHROMOSOMES

divisions; and in that paper I have referred to the observations of other workers cor-
roborative of the individuality of the chromosomes, so that they need uot be recalled here.
In such Hemiptera as the Coreidce there is no true rest stage in the growth period, so that
definite chromosomal outlines can be determined throughout the growth period. And
even in those Hemiptera where a rest stage does occur in this period, in which the
chromosomes become reticular and practically indistinguishable from one another, all
evidence renders it probable that the chromosomal individuality is retained through this
period i. e., that a particular univalent chromosome of the maturation mitosis represents
a particular one of the spermatogonia. This evidence is as follows : The chromatin
nucleoli, which are only modified chromosomes, retain their compact form and so are
readily distinguishable throughout all periods of these generations. Then in those cases
where there is an uneven number of chromosomes in the spermatogonia, and the odd one
remains univalent in the spermatocytes, this odd one can always be distinguished in the
maturation mitoses. Then where two of the spermatogonic chromosomes are particularly
large, there is always found in the first maturation mitosis one particularly large bivalent
chromosome which can only correspond to those two. Further, after a rest stage there is
always the same number of chromosomes as were present before that rest stage. All this
evidence speaks for a chromosome of a spermatogonium corresponding to a chromosome
of a spermatocyte ; and if in these generations there is a maintenance of chromosomal
individuality, there is a probability that there is such a maintenance through all
generations of the germinal cycle. But this conclusion by no means implies that a
chromosome of one generation is actually the same as a chromosome of another. For
we know that each chromosome is halved in an equation division, and that each daughter
chromosome so produced must increase to a volume equal to that of the mother chromo-
some before it enters upon a second mitosis. Thus new substance must continually
be elaborated by the chromosomes during the rest stages, and in the course of this
elaboration the old substance of the chromosome and its physical form are correspond-
ingly changed. There is no evidence that chromosomal substance remains unchanged
from generation to generation, for all evidence shows that it undergoes metabolic change
and growth in the rest stages. But nevertheless it seems very probable that a chromo-
some of one generation is a derivative of a particular chromosome of the preceding genera-
tion, and that the chromosomes may thus be said to maintain themselves as entities
through successive generations.

There are many cases where chromosomal boundaries are indistinguishable in the
rest stages, so that in such cases it has been argued that the chromosomes show no indi-
viduality. But these are negative examples, and positive cases that speak for the main-
tenance of the chromosomal individuality must be considered as the decisive ones. No

OF THE GERM CELLS OF METAZOA. 201

other assumption can so well explain the maintenance of a constant number. The fact
that the chromosomes build up new substance at one stage, and give off waste
products at another, does not invalidate our conclusion. Those who deny the maintenance
of chromosomal individuality, on the basis of a study of objects where the chromosomes
do not appear to be continuous from generation to generation, are not justified in conclud-
ing that there is never a maintenance of the individuality until they have examined the
positive cases. And since there have been shown to be positive cases, we must conclude
(1) either that in all normal mitoses of the germinal cycle the chromosomal individuality
is maintained, or (2) that it is preserved in some cases but not in, others. The fact that
it has been demonstrated for some cases renders it probable that, is maintained in all
cases, even though it cannot be demonstrated in all.

2. The chromatin nucleoli.

The term " chromatin nucleolus" was applied by me (1898) to the remarkable
chromatin element, in form like a nucleolus but in behavior and staining reactions like
a chromosome, in the nuclei of spermatocytes of Euchistus (Pentatoma) ; the name was
given in order to denote this double similarity, though I fully realized that this structure
was in all essentials a modified chromosome. A special monograph (1899 b) was devoted
by me to the true nucleolus (plasmosome, Ogata) of Metazoan cells ; and in a lecture at
the Woods Holl laboratory (1898 b) I classified the other "nucleolar" structures as
" karyosomes, which are merely thickened nodal points of the chromatin reticulum ;
further, what I shall term the ' chromatin-nucleus ' [typographical error for ' chromatin
nucleolus'], which is found in certain spermatocytes; and then various structures
which stain neither like the true nucleolus nor the chromatin, and to which such terms
as ' Paramcleoli,' ' Nebennucleoli,' and ' Pseudonucleoli ' have been applied. It is
one of the most difficult questions to determine the nature and correspondence of the

latter structures But from the cases studied by me, it would appear that

some of these structures in Metazoa probably must be placed within the category of true
nucleoli, and be regarded as true nucleoli of a different chemical nature ;.... our
criterion of nucleoli probably should not be based as much upon chemical as morpho-
logical facts."

The chromatin nucleolus is a modified chromosome, as both my earlier and the
present observations show. The term may have been injudiciously chosen, since I myself
showed that it has nothing in common with a true nucleolus except sometimes in the
form and in containing vacuoles. In its stead McClung (1899) has given the name
" accessory chromosome," and Paulmier (1899) the name "small chromosome." But
McCiung's term is not satisfactory in not being at all definite, and Paulmier's term is not

202

MONTGOMERY A STUDY OF THE CHROMOSOMES

applicable to those cases where it is as large as the other chromosomes. Hence I con-
sider that confusion in terminology would be best avoided by retaining my original term,
" chromatin nucleolus."

In my first description of the spermatogenesis of Euchistus (1898 a), I stated that
I could not find chromatin nucleoli in spermatogonia, and that one appeared for the
first time in the spermatocyte by a metamorphosis of one of the fourteen chromosomes ;
this was an error that I have corrected in the present paper, for in Euchistus, just as
Paulmier (1899) correctly described for Anasa, there are two chromatin nucleoli in the
spermatogonium, and these unite in the spermatocyte to form one bivalent one. And in
all the Hemiptera examined by me the larger chromatin nucleoli of the spermatocytes
are derivatives of chromatin nucleoli of the spermatogonia, except the remarkable
" chromosome x " of Protenor, to which we shall return. As far as I have been able to
determine, the chromatin nucleoli are always halved in the mitoses of the spermatogonia.

(a) General Characteristics.

The chromatin nucleoli are morphologically chromosomes, undergoing division in
mitosis like the other chromosomes, but differing from them in the rest stage by preserv-
ing a definite (usually rounded) form. There is also another difference which is of great
use in their study : by the use of the double stain of Hermann, saffranine and gentian
violet, the chromosomes proper stain red only in mitosis and violet in the rest stage,
while the chromatin nucloli stain red in the rest stage also, and so can be sharply dis-
tinguished from the chromatin of the chromosomes.* Thus the chromatin nucleoli of
the Hemiptera seem to retain at all stages the stain characteristic for the substance of the
chromosomes when in the height of mitosis. In the Hemiptera examined the true
nucleoli never takes this red stain, but take the violet, so that they may in this way be
easily distinguished from the chromatin nucleoli.f With iron-hrematoxylin staining
the chromatin nucleoli stain more intensely than the chromosomes in the rest stage ; but
with this stain the true nucleoli generally stain deep black like the chromatin nucleoli,
so that it is far less satisfactory than the preceding method for differentiating these two
structures. With the triple stain of Ehrlich-Biondi-Heidenhain, the chromatin nucleoli

* The only exception to this staining reaction was found in the spermatocytes of Hygotrechus, where in the rest
stage the chromatin nucleoli always take the violet stain.

t However, in cells of many other Melazoa I have found that the true nucleoli show a particular eleetivity for
the saffranine, so that chemical reactions are not tests for true nucleoli ; nucleoli may differ chemically from one
another, even in the same cell at the same stage, or at different stages, and no better case of this may be mentioned
than the ovocytes of the growth period of Gryllui. In the Hemiptera the true nucleoli are generally much larger
than the chromatin nucleoli, more or less irregular in outline, and they usually occupy a more or less central position
in the nucleus (though I have mentioned two or three exceptions), while the chromatin nucleoli in the spermato-
cytes are generally in contact with the nuclear membrane.

OF THE GERM CELLS OF METAZOA. 203

and chromosomes in the rest stage are green, the true nucleoli red ; with Delafield's
hgematoxylin and eosin, the chromatiu. imcleoli and chroinosoines stain blue, the true
nucleolus red ; with both these methods the chromatin nucleoli stain more intensely
than the chromosomes in the rest stage.

In the spermatogonia the chromatin nucleoli are generally small, often very minute
and difficult to distinguish. They are apparently always halved in the spermatogonic
mitoses, though I have not been able to determine this for all species. They always
increase in volume during the growth period of the spermatocytes ; the relative amount
of this increase varies in different species, but it is a number of times greater than in the
spermatogonia. During this increase in mass there can be found in most cases a clear
vacuole in the chromatin nucleolus, so that the increase would appear to be not so much
one of the proper substance of the chromatin nucleolus as of an intussusception of fluid
from without ; I have not figured these vacuoles in the present paper, but showed them
in a preceding one (1898 a). In the prophases of the first maturation division the
chromatin nucleolus decreases very considerably in volume, until when the monaster
stage is reached its volume may be very little greater than in the spermatogonia. It is
possible that its increase in mass during the growth period may be due in part to a decon-
densation of its substance, but in some part at least it is due to the above-described
taking in of fluid substance from without.

In the spermatogonia they are irregular in position, sometimes close together, some-
times separated, but usually not in contact with the nuclear membrane ; in these cells
they are very frequently in all species apposed to the true nucleoli.* In the growth
period of the spermatocytes they are more regular in position : thus they are separated
from the true nucleoli and then always in contact with the nuclear membrane in
Euchistus, Mormidia, Peribalus, Cosmopepla, Nezara, Brochymena, Trichopepla, Eury-
gaster, Metapodius, Anasa, Chariestents (generally), Corizus, Harmostes, Calocoris, Hygo-
trechus, Limnotrechus ; they are, as a rule, apposed to the true nucleoli in Podisus,
Perillus, Ccenns, Alydm, Protenor, Peliopelta (not always), (Edancala, Oncopeltus (not
always), Leptoptema, Pcecilocapsus, Phymata, Coriscus, the Reduviidce, Limnobates and
Zultha. We shall refer again to the significance of this apposition. A chromatin
nucleolus and a true nucleolus closely attached together constitute a " double nucleolus ;"
it remains to be shown whether the " double nucleoli " of certain cells of other Metazoa,
as e. g. in the cells of Sertoli of Salamandra and Mus, as described by Hermann and
others, may also be cases of apposed chromatin nucleoli and true nucleoli. It is very
characteristic of the chromatin nucleoli of the spermatocytes in the growth period to be
closely apposed to the nuclear membrane. And when they are apposed to true nucleoli,

* In the spermatogonia there are in all species an irregular number of true nucleoli, but in the spermatocytes of
most of the species examined there is regularly one large one.

204 MONTGOMERY A STUDY OF THE CHROMOSOMES

they retain this position if they are of comparatively large size (as in most Pentatomidce)
but do not when they are of comparatively small size (as particularly in the Reduviidce).
Now the true nucleolus is rarely in contact with the nuclear membrane when it is not
apposed to a chromatiu nucleolus, so that when the two are mutually apposed it appears
to depend upon their relative volumes whether they will be peripheral or central iu
position a very large chromatiu nucleolus pulling to the periphery the true nucleolus, a
small chromatin nucleolus being pulled by the true nucleolus toward the centre of
the nucleus.

That the chromatin nucleoli are morphologically chromosomes is shown particularly
in mitosis, when they simulate in form and divide like chromosomes; an examination of
the chromatin nucleoli, designated N. 2, in the plates of the present paper, demonstrate
this point. The univalent chromatin nucleoli of the spermatogonia generally unite to
form bivalent ones in the synapsis stage just as do the chromosomes, and generally the
number of them in the spermatocytes is just half of that in the spermatogonia (certain
exceptions shall be considered later). In the first maturation division each bivalent
chromatin nucleolus is halved transversely (reduction division) just as are the bivalent
chromosomes.

Thus the chromatin nucleoli are essentially chromosomes, but chromosomes which
preserve a compact form and dense structure throughout the rest period.

(b) Number and Valence.

It is most frequently the case that there are two uuivalent chromatiu nucleoli in the
spermatogonia, as in the Pentatomidce, Euryyaster, Coreidce, Peliopelta, (Edancala, Phy-
mata, Coriscus (?), Hygotrechus, Zaitha. In these cases the two chromatin nucleoli join
together to form one bivalent one in the growth period ; but in Euchislnx tristigmus and
Chariesterus the two frequently remain separate in the growth period, yet this is not a
complete separation for there would seem to be a linin connection between the two, since
they generally become more closely approximated at the time of the first maturation
division when such a linin connection may be seen. But in some of these species (Peri-
balus, Coenus, Trichopepla, Coriscus) besides the bivalent one are found one or two (three
or four in Trichopepla) much smaller ones, which appear to be univalent, in the sperma-
tocytes ; whether these are represented in the spermatogonia or whether they arise in the
spermatocytes for the first time, I could not determine on account of their minuteness.

In Oymus and Jchnodemus I could not determine the number in the spermatogonia,
but since here there is one bivalent one in the spermatocytes, there would probably be
two univaleut ones in the spermatogonia. The same would probably be true for Corizus
and Leptopterna.

The Reduviidce and certain Capsidat show a larger number of chromatin nucleoli.

OF THE GERM CELLS OF METAZOA. 205

Thus in Acholla there are eight in the spermatogonia which form four bivalent ones in
the spermatocytes ; in Sinea and Milt/as there are three bivalent ones in the spermato-
cytes, and accordingly probably six in the spermatogonia ; in Prionidus there are appar-
ently five in the spermatogonia, two of which unite in the spermatocytes to make one
bivalent one while the three others remain univalent. In Calocoris there are in the
spermatocytes two bivalent and one univalent ; in Poeciloeapsm lineatus, one bivalent and
one apparently univalent, ; in P. goniphorus certainly four, and possibly five, bivalent
chromatin nucleoli in the spermatocytes. In these forms one or more of the chromatin
nucleoli may have their component parts more or less widely separated, but these sepa-
rated components generally come together before the first maturation mitosis.

Bivalent chromatin nucleoli which have their components in close mutual apposition
appear always to be transversely divided in the first maturation division (reduction divi-
sion); for the behavior of those which are univalent I refer to the chapter on " Observa-
tions."

Oncopeltus affords the interesting case where the two chromatin nucleoli of the
spermatogonia are apparently each bivalent bivalent elements in a generation where all
the chromatin elements are usually supposed to be univalent.

(c) The Peculiar Chromosome of Protenor.

In Protenor for the details compare the description in the chapter on " Observa-
tions " there are two chromatin nucleoli in the spermatogonia, which in the spermatocytes
unite to form a bivalent one, as we have just seen to be quite generally the rule in Hemiptera.
But the largest of the chromosomes of the spermatogonic mitoses, the "chromosome x,"
does not behave in the growth period of the spermotocytes like the other chromosomes,
but is similar to a chromatin nucleolus in preserving a compact form and in retaining the
saffranine stain. This is the odd chromosome, the eleventh, which does not combine with
any other during the synapsis stage, and which cannot be distinguished in the rest stage
of the spermatogonia because there it behaves exactly like the other chromosomes, and
takes part in the formation of the nuclear reticulum. This is the only case in the
Hemiptera, where one chromosome becomes differentiated into a chromatin nucleolus for
the first time in the spermatocyte generation, unless the minute chromatin nucleoli of
Peribalus, Ccenus, Trichopepla and Coriscus may be found to have a similar history.

(d) Function.

All the observations show that the chromatin nucleoli are modified chromosomes,
which behave essentially like the chromosomes in mitosis but quite differently in the rest
stage.

Paulmier (1899) has suggested that they are degenerate chromosomes. This would

206 MONTGOMERY A STUDY OF THE CHROMOSOMES

seem to be true to some extent, but not wholly correct for the following reasons. In a
large number of the species examined the chromatin nucleoli are regularly closely
apposed to the surfaces of true nucleoli. Now it seems probable that the true nucleoli are
masses of substances formed by the metabolism of the cell, probably waste substances
(Montgomery, 1899 b). When we find accordingly the mutual apposition of them to
chromatin nucleoli, it would be permissible to conclude that the chromatin nucleoli are
chromosomes which are especially concerned with nudeolar metabolism. And this I
think would be the correct interpretation. The chromatin nucleoli are in that sense de-
generate, that they no longer behave like the other chromosomes in the rest stages ; but
they would appear to be specialized for a metabolic function. Thus it might be that in
the Insects the chromatin nucleoli are those chromosomes which either exert a greater meta-
bolic activity than the other chromosomes, or which carry out some special kind of meta-
bolism ; and from this point of view they would certainly seem to be much more than
degenerate organs. As to their origin, compare the chapter on the " Number of Chromo-

somes."

Like the chromosomes, the number of chromatin nucleoli in the germ cells appears to
be a fixed one for the species. In somatic cells they are often more numerous than in the
germ cells, however, but that somatic difference will not be discussed in this paper which
concerns itself with the germinal cycle.

(e) Occurrence.

Chromatin nucleoli were found by me in all the Hemiptera examined, and I have
found them also in Coleoptem (Harpalus) and Orlhoptera (Gryllm, Ceutophilm).
McClung (1899, 1900) has described them in various Orthoptera, in some forms of which
they are larger than the chromosomes. Finally, my student, Miss Wallace (1900), has
found them in the sperm atogenesis of a spider (Agalenid).

Accordingly they would seem to be present in the Insects and Arachnids, but are
apparently absent in the Crustacea, and I have shown (1901) that they are not present in
Peripatus.

It is, however, quite possible that they will be discovered in other forms, if proper
attention is paid to them. The question of their ontogenetic origin will be considered
later (compare the section on the " Significance of the uneven number " of chromosomes).

3. The Number of Chromosomes.

The following table shows the number and valence of the chromosomes and chroma-
tin nucleoli in the Hemiptera examined, all of which has been explained in detail in the
chapter on " Observations." The abbreviation " univ" has been employed for " univa-
lent," and " biv." for " bivalent."

OF THE GERM CELLS OF METAZOA.

207

Sl'KCIF.S.

<

SPEBMATOGONIA.

SPERMATOCTTES.

Chromosomes.

Chromatin
nuclcoli.

Chromosomes.

Chromatin
nuclcoli.

No.

Valence.

No.

Valence.

No.

Valence.

No.

Valence.

Eucliistus variolarius

14
12
14
14
14
1C
14
14
14

12
14

univ.
iiniv.
univ.
univ.
univ.
univ.
univ.
univ.
univ.

univ.
univ.

2
2
2
2
2
2
2
2
2

2
2

univ.
univ.
univ.
univ.
univ
univ.
univ.
univ.
univ.

univ
nniv.

7
6

7
7
7
8

biv.
biv.
biv.
biv.
biv.
biv.

1
Ior2
1
1

1
1

biv.
biv. or nniv.
biv.
biv.
biv.
biv.

E. tristigmus

Podisus spinosus

Mormidea lumens

Peribalus lirnbolaris

Cosmopepla carnifex

Brochvniena sp

7
7

C

7

6

10
10
10
10
12
6

6
G
6
5

12

7
7

G

7

ia

14

16?

17? 14?
14

9

12
14

biv.
biv.

biv.

biv.

biv.

biv.
biv.
biv.
biv.
biv.
biv.

fSbiv. \
\ 1 univ. /

biv.

f 5 biv. 1
\1 univ. /

biv.

biv. (all?)
biv.
biv.

f 5 biv. 1
\ 1 univ. J

biv.
biv.

biv.

biv.

hiv.
biv.

biv.

biv.
biv.

1
1

2

4 or 5

1
1
1
1
1
1
1

1

2 or 3
1
2

1
2 or 3
1

1

2

1

3

2

4
1

2

4
3

biv.
biv.

f 1 biv.
\1 univ.

f 1 biv.
\8or4univ.

biv.
biv.
biv.
biv.
biv.
biv.
biv.

biv.

f 1 biv.
\1 or2uuiv.

biv.
biv.

biv.
?
biv.

biv.

biv.
biv.

f 2 biv.
\1 uuiv.

f 1 biv.
1 1 univ.

biv.
biv.

f 1 biv.
\2 univ.

biv.

biv.

Perillus confluens

Cceuus delius . . . .

Eurygaster alternatus

Auasa tristis

20
20
20
20

univ.
univ.
univ.
univ.

2

2
2
2

univ.
univ.
univ.
univ.

A. SD - .

Metapodius terminalis

Alydus pilosulus.

12

11

univ.
univ.

2
2

univ.
univ

A. eurinus

Corizus lateralis

ITannostes reflexulus

11
11

univ.

/ 10 univ. \
\1 biv. /

2
2

univ.
univ.

Protenor belfragei

Uyuius angustatus

Ichnodemus falicus

14
14

11
14

univ.
univ.

univ.
univ.

2

2

2
2

univ.
univ.

univ.
biv.

Peliopelta abbreviata

G'Mancahx dorsalis

Oncopeltus fasciatus .

Leptopterua dolabrata

Calocoris rapidus

28?

univ.

5?

univ.

Poecilocapsus lineatus

P. goniphorus

Phymata sp

28

univ.

2

univ.

Coriscus ferus

Acholla multispinosa

24

univ.

8

univ.

Sinea diadema

208

MONTGOMERY A STUDY OF THE CHROMOSOMES

SPECIES.

SPKKMATOGONIA,.

SPEKMATOCYTES.

Chromosomes.

Chromatin
nucleoli.

Chromosomes.

Chromatin
nucleoli.

No.

Valence.

No.

Valence.

No.

Valence.

No.

Valence.

21?

1

5?

univ.

Milyas cinctus . .

3
1
2

1?

biv.
biv.
univ.
biv.?

Hygotrechus sp

18

univ.

2

univ.

9

10?

biv.

biv.

20?
22

Zaitha sp. .......

univ.

2

univ.

11

biv.

2

univ.

The following general deductions may be drawn from the consideration of these
facts. Whenever there is an even number of univalent chromosomes in the spermato-
gonia, they unite in the synapsis to produce exactly half this number of bivalent
chromosomes. When there is an uneven number in the spermatogonia (Alydus eurinus,
Harmostes, Protenor, (Edancala) all but one of the chromosomes unite in the synapsis
to form bivalent chromosomes, while the odd one remains single.

(a) Number and Genetic Relationship.

At the outset of the present study I was particularly interested to determine
whether the numbers of chromosomes might afford clues to the relationship of the group
of the Hemiptera heteroplera ; that is, to learn, if possible, whether the number would
afford a taxonomic criterion. Most of the families of the Hemiptera are very rich in
species, however, and I have been able to procure only a few species for study, so that
the present beginning must be continued on many more species before .any conclusion
can be reached.

In the Pentatomidce, counting the chromosomes in the spermatogonia, and not
including the chromatin nucleoli, we find the number varies between twelve and sixteen,
fourteen being most usual ; in the Coreidoe we find twenty in Anasa and Metapodim, in
Alydus eleven and twelve, in Harmostes and Protenor eleven. In the Lygceidw, fourteen
in Ichnodemus, Peliopelta and Oncopeltus, eleven in (Edancala, and probably twenty-
four in Cymus. These were the three families of which the most species were examined.
From these numbers it will be seen that there is considerable variation in number for
the different species of one and the same family. Accordingly we must conclude either
(1) that the number of chromosomes is easily modified and changed, so that it has little

OF THE GERM CELLS OF METAZOA. 209

taxonomic value ; or (2), that the families of the Heiniptera heteroptera, as they are at
present defined, are artificial and not natural groups. I would incline to the latter view,
since all our facts would show that chromosomes are very conservative structures, and
that the germinal cycle is conservative ; probably the soma may be modified by the
action of the environment to considerable extent, before any such action would produce
a change in the number of chromosomes. If this standpoint is correct, then the number
of the chromosomes would be a very important consideration in deciding the relationship
of species ; thus the Coreidce would have to be subdivided into a sub-group with twenty
to twenty-four chromosomes (Anasa, Metapodius, Chariesterus), and into one with eleven
or twelve chromosomes (Alydus, Corizus, Harmostcs, Protenor}. But it would be a
reductio ad absurdum to say that all forms with twelve chromosomes must be related, or
all forms with twenty ; the relative boundaries of a family must still be determined from
the standpoint of broad comparative anatomy, and then within a group so defined the
chromosomal number might be used as a basis for further subdivision.

(b) Factors Determining the Number.

A problem of great importance, and one that would seem to lie close to the root of
all nuclear phenomena, is that concerning the factors which determine the number of
chromosomes. The germ cells of each species have a fixed number of chromosomes, but
different species show a very different number, from Ascaris megalocephala univalens
with 2 up to Artemia with about 180. What is it that determines this numerical differ-
ence? A consideration of the various thinkable factors allows us to limit the problem
somewhat, by excluding certain ones which are not real factors. Here we may consider
in what relation to chromosomal number stand centrosomes and achromatic spindle
elements, number of nucleoli, mass of nucleus and cell body, form of cell, volume and
form of chromosomes.

The size, number and specific peculiarities of the centrosomes seem to have no connec-
tion with the number of the chromosomes. The definite number of chromosomes appears
in the prophases of mitosis while the nuclear membrane is still intact, and when the cen-
trosomes are only commencing to exert an influence upon the other cell constituents.
Even the longitudinal splitting of the chromosomes would appear to be an automatic move-
ment on their part. The centrosomes may well be centres of movements which produce
the separation of the daughter chromosomes, but there appears to be no correlation
between the centrosomes and the chromosomal number. And this seems also to be the
case with regard to the spindle fibres. Central spindle fibres and polar radiations may
vary in their phenomena in different generations of the same species, but the number of
the chromosomes remains constant in all generations of the germ cells, for the apparent

210 MONTGOMERY A STUDY OF THE CHKOMOSOMES

halving of the number just before the maturation mitoses is not a real halving of the
normal number, but only a grouping into pairs. And the mantle fibres, those which
connect the centrosomes with the chromosomes, seem in fact to have' their number deter-
mined to some extent by the number of the chromosomes, and not to determine that number,
for the definite number of chromosomes appear in mitosis before the mantle fibres arise.
The number of the connective fibres, those whieh connect corresponding daughter
chromosomes in the anaphase, is certainly determined by the number of the chromosomes,
for these fibres are stretched-out portions of the linin matrices of the chromosomes.
Thus the centrosomes and the achromatic spindle structures may play a part in the dis-
tribution of the chromosomes, but apparently have no part in the determination of their
number.

As for the true nucleoli, their number, volume and position seem to be in no way
regulative of the chromosomal number. The nucleolar number in one species is gener-
ally variable, and correlatively also the volume and position, while the chromosomal
number is constant ; in the paternal germ cells the number of nucleoli is frequently
different (and generally smaller) than that in the maternal cells of a species, but in both
kind of cells the number of chromosomes is the same.

The mass of the nucleus or of the cell body, or the relative mass of the two, might
seem a priori to stand in connection with the chromosomal number, yet an examination
of the facts shows this is probably not the case. For instance, in one species the huge
ovocytes and the much smaller spermatocytes have the same number of chromosomes, and
a small ovogonium has the same number of chromosomes as a large ovocyte. And in the
case of the ovocyte the volume of the nucleus is relatively small, in the spermatocyte rela-
tively large in proportion to the mass of the cell body, yet the number of the chromo-
somes is the same. The mass of the chromatin substance may be more or less proportion-
ate to the volume of the nucleus, but the number of the chromosomes appears not to be ; the
large number of chromosomes in Artemia (as determined by Brauer) is not correlated
with a large nucleus ; and as the figures of the present paper show, cells of approximately
the same size from different species may show very different chromosomal numbers. The
form of the cell is regulated to great extent by external influences, and variations in the
form produce no modification of the chromosomal number.

The volume of the chromosomes in mitosis is dependent upon their number, since
the volume of the chromatin stands in more or less direct ratio to the volume of the
nucleus. The form of the chromosomes is more or less dependent upon their number,
inasmuch as long ribbon-shaped chromosomes occur only where there are a small number,
and rounded ones where there are a larger number present. Yet in all the Hemiptera,
with their considerable differences in chromosomal number, the form of the chromosomes

OF THE GERM CELLS OF METAZOA. 211

remains quite constant, each univalent chromosome being generally slightly elongate.
Thus the form is not wholly dependent upon the number, but particular groups of
Metazoa appear characterized by particular forms of chromosomes, as e. g. the Mollusca
with their rod-shaped chromosomes. Then the form of bivalent chromosomes is
dependent upon the mode of junction of the component univalent elements. In no
sense, however, can the form of the chromosomes be said to determine their number.

So far our considerations lead to only negative results; centrosomes and achromatic
spindle structures, nucleoli, absolute and relative mass of nucleus and cytoplasm, cell form
and chromosomal form seem not to be factors determining the number of chromosomes.
As I attempted to show in a preceding paper on Pei-ipntus (1001), the chromosomes must
be regarded as individuals of a lower grade than what I termed the " nuclear element,"
namely, the linin spirem with the chromosomes arranged upon it. The problem is really,
then, why does this nuclear element show in one species a certain number of chromatin
segregations, in another species a different number ? The more recent chemico-physio-
logical studies would tend to show that the chromosomes are centres of metabolic activity,
and accordingly the problem of the factors governing the chromosomal number may be
closely connected with the phenomena of metabolism ; the number of the chromosomes
may be dependent upon the nature of the metabolism, as upon either the chemical nature
of the chromosomes themselves or upon that of the cell nutriment. The latter might be
experimentally tested by changing the food of a species, and observing whether differ-
ences in the number of the chromosomes might thereby be obtained. But to conclude
that the number of the chromosomes is dependent upon the nature of the metabolism
docs not solve the problem but only states it more precisely.

Another question which arises in this connection is whether a small or a large
number of chromosomes is to be regarded as the primitive condition. A priori it woidd
appear probable that at an early phyletic period the number of chromosomes was not
fixed for the species but variable, and that by a process of natural selection the number
gradually became fixed. But as a species gradually changes into another form the num-
ber of chromosomes may also be changed, as will be shown in the next section, so that
we may speak of an evolution of the number of chromosomes. On the principle of the
law of greater condensation of organs in progressing evolution, it might be that a large
number of chromosomes represents a more primitive condition than a smaller number.
Within such a group as the Hemiptera heteroptera, for instance, forms like the Belosto-
matidce, Reduviidce, Capsidce and Phymatidce would be primitive in possessing a larger
number of chromosomes, while the Pentatomidce, Scutellariidce and Lygceidce in possess-
ing a smaller number should be regarded as more specialized more highly developed.
From such a standpoint as this, the chromosomal number would be of taxonomic

212 MONTGOMERY A STUDY OF THE CHROMOSOMES

value it would be a signboard of degree of specialization within the group. From
such a point of view it might even be possible to construct a cellular classification which
would Iiavf great value in that it would employ truly conservative structures. The cen-
trosomes and central spindles have been considered phyletically in this way by Biitschli,
Lauterborn, Heidenhain, Calkins and others, but so far the chromosomes have not been
considered from such a standpoint, although they in many respects appear more conserv-
ative than centrosomes and achromatic spindle structures.

Accordingly, though the present study has not given a solution to the problem of
the factors governing the number of chromosomes, except in showing that it must be
sought in the phenomena of metabolism, yet it would show that chromosomal number may
be employed as a criterion of relationship if it be used with caution and with due consid-
eration of a broad comparative treatment of other. structures. And the reason is because
the chromosomes seem to be highly conservative, their number constant for the species,
and because in a certain sense they -represent the most important vital structures. We
should not conclude that all forms with e. g. ten chromosomes should be ranked as closely
related ; all with e. g. twenty-four as composing another natural group. But within a
certain group which has been defined on a broadly comparative basis such a group as the
Hemiptcra hcteroptcra for instance, the chromosomal number would perhaps be a clue
to the relative degree of specialization of the species.

(c) Significance of the uneven normal number.

One of the most unexpected results of this investigation was the discovery that in
some species there is an uneven number of chromosomes in the spermatogonia, i. e., an
uneven normal number, whereas heretofore all observation and assumption has been that
the normal number is always an even one. Of the Hemiptera studied, four species show
an uneven number of chromosomes in the spermatogonia, namely, Alydus eurinus, Har-
mostes reflexulus, Protenor belfragei and (Edancala dorsalis ; in all of these the number
is eleven, and in the synapsis stage one chromosome remains univalent while the ten re-
maining combine to form five bivalent chromosomes. AVhat is the significance and origin
of this uneven number?

Now, as far as our facts go, it seems that the number of the chromosomes is constant
for the species, and that the paternal and maternal germ cells of a species have the same
number ; this appears to be one of the points in the correspondence of the ovogenesis and
spermatogenesis first determined by Henking (1890) and O. Hertwig (1890). Whenever
there is this correspondence in number and valence, then in fertilization, when the
paternal chromosomes- are added to the maternal, an even number of chromosomes
should result, and if the chromosomes maintain their individuality through the succeed-

OF THE GERM CELLS OF METAZOA. 213

ing generations, there should be an even number in the spermatogonia and ovogonia. The
uneven number discovered in the four species mentioned may have arisen in one of two
ways : through bastardization, or through a mitotic abnormality, each of which possibilities
may now be considered.

In the case of bastardization of a germ cell with one chromosomal number by a
germ cell from another species with a different number, the uneven number eleven might
be secured if a paternal germ cell of a species A, with the normal chromosomal number
twelve, fertilizes a maternal germ cell of a species B with the normal number ten. In
species A the reduction in number of the chromosomes would give six univalent chromo-
somes in the spermatid, and in species B five univalent chromosomes in the matured
ovum ; conjugation of the two cells would then result in 6 + 5 = 11 univalent chromo-
somes. In this way an uneven chromosomal number may have arisen by the conjugation
of the germ cells of species with different numbers of chromosomes. But the objection to
this view lies in the fact that hybridization of distinct species is generally infertile, and
species which would have different chromosomal numbers would probably be quite distinct.

More probably, then, the uneven number may have originated through abnormalities
of mitosis, and there would be many possibilities for such an occurrence. (1) It may have
arisen in a spermatogonic mitosis, in a species where the ancestral normal number is
twelve, by the chromatin of the spirem segregating abnormally into only eleven chromo-
somes, so that one of these chromosomes would be virtually bivalent. This would seem
to have been the origin of the large odd chromosome x of Protenor be/fragci, which ap-
pears bipartite in the spermatogonia and also in the spermatocytes, though in the latter it
does not conjugate with any other chromosome. Such a case would be a deficiency in the
segregation of the chromatin in the prophases of mitosis. (2) Or the uneven chromo-
somal number may have arisen by an unequal distribution of the chromosomes in the
anaphases of mitosis, so that the daughter cells would not receive equal numbers.

Now, whether the uneven chromosomal number had originated through bastardiza-
tion, or, what would be more probable, through abnormality in mitosis, it is interesting to
determine how this number can perpetuate itself through different generations of the
species, for my observations show that in the species where it is found it occurs in all
individuals. The following table gives in condensed form the mode of reduction of the
chromosomal number and valence in the four species in question, the chromatiu nucleoli
being omitted for the sake of simplicity :

214

MONTGOMERY A STUDY OP THE CHROMOSOMES

SPECIES.

SPERM ATOGONIUM.

FIBST SPERMATOCYTE.

SECOND SPEHMATOCYTE.

SPEKJIATID.

Alytlus eurinus.

11 univalent.

5 bivalent.
1 univalent.

5 univalent.
1 semivalent.

5 or 6 semivalent.

Ilarmostes reflex-
ulus.

11 univalent.

5 bivalent.
1 univalent.

5 univalent.
1 semivalent.

5 or 6 semivalent.

Protenor bel-
fragei.

10 univalent.
1 bivalent.

6 bivalent.
1 bivalent.

5 univalent.
1 univalent.

5 semivalent.
1 or univalent.

CEdancala dor-
salis.

11 univalent.

5 bivalent.
1 univalent.

5 univalent.
1 semivalent.

5 semivalcnt.
1 (orOV) semivalent.

It may appear strange that the chromosomes of the spermaticls are classed as semi-
valent since they are generally considered univalent ; as I have explained in an earlier
portion of this paper, however, they must be regarded as semivalent on account of tin-
absence of a rest stage between the maturation mitoses, though they are virtually univa-
lent on account of their increase in mass during the growth period. Protenor lrlfr<nj> i
differs from the three other species in showing a bivalent chromosome in the spermatogo-
nium, which chromosome is consequently bivalent in the first spermatocyte even though
it unites with no other during the synapsis stage. All four species have in common the
phenomenon that the odd chromosome does not conjugate with any other during the syn-
apsis stage, but remains separate. In Alydus eurinus, Harmostes reflexulm, and Protenor
belfragei this odd chromosome does not divide in the second maturation mitosis, but passes
undivided into one of the two spermatids. In (Edancala dorsalis I was unable to deter-
mine its behavior in this mitosis, though I have no reason to suppose that here it behaves
differently from the other species. This unequal distribution of the odd chromosome in
the second maturation mitosis is evidently in some way dependent upon its not having
united with a fellow-chromosome during the preceding synapsis stage. What concerns us
particularly at present is the fact that in these species with an uneven normal number of
chromosomes, unlike those with an even number, one chromosome (the odd one) is not
divided in the second maturation mitosis, but passes undivided into one of the daughter
cells (spermatids); half of the spermatids then have six chromosomes and half have only
five.

Bearing this point in mind, let us see how the uneven chromosomal number may be
perpetuated from individual to individual. This may be occasioned by one of two possi-
bilities. (1) The paternal germ cells having eleven chromosomes in the spermatogonia
and either five or six in the spermatids, there is the probability that the maternal germ
cells (ova) may have a corresponding number of chromosomes. If a spermatozoon with

OF THE GERM CELLS OF METAZOA. 215

five (or six) chromosomes conjugates with an ovum with five (or six), so that each of the
conjoints has the same number, an even number would result in the fertilized ovum. But
if a spermatozoon with five (or six) chromosomes unite with an ovum with six (or five),
the conjoints Jiaving then different numbers of chromosomes, the fertilized ovum would
have the uneven number eleven ; the uneven number would then be perpetuated from
individuaj to individual, so long as the conjugating cells have different numbers of chro-
mosomes, and so long as the odd chromosome does not divide in one of the maturation
mitoses. (2) Or germ cells from individuals with an uneven normal number of chromo-
somes, by conjugating with germ cells from individuals with an even number, would occa-
sion an uneven number in the fertilized ovum. Either of these possibilities would suffice
to explain the transference of the uneven number from individual to individual, though
the first possibility would appear the more probable.

So far we have considered the origin of the uneven chromosomal number and the
mode by which it is perpetuated from individual to individual. We have now to discuss
its significance. Most of the Hemiptera examined by me show an even normal number of
chromosomes ; only four showed an uneven number, and in no other Metazoa has an un-
even number, to my knowledge, been found. The uneven number would accordingly
appear to be unusual. It seems to me probable that the uneven number represents a
transition stage between a higher number and a lower, or the converse, and it is unusual
because the transition stage is probably shorter than the earlier and the later stages. The
number of the chromosomes varies quite considerably in the different species of the Hemip-
tera heteroptera, but we cannot suppose that the number was constant for each species
from the beginning any more than we can consider that the species have always remained
unchanged ; there must have been an evolution of the chromosomal number, as there has
been of the species. It is quite possible that an even number of chromosomes, as e. g.
twelve, may have changed into an even number (ten) without first passing through the stage
of the uneven number (eleven). This might take place by the number ten appearing
simultaneously in both paternal and maternal germ cells through some abnormality or
deficiency in mitosis. But it is far more probable that such a mitotic abnormality would
not occur coincidently in both kinds of cells more probable, e. g., that a paternal germ
cell, acquiring an abnormal number of chromosomes by some fault in the process of mitosis,
would conjugate with a maternal germ cell with the normal number; the result of such
a union would be of course an uneven number. On this argument, when the chromo-
somal number changes, the period of change would be characterized by an uneven num-
ber of chromosomes. Ultimately an even number of next lower or next higher order
would be reached, and that number must persist longer than the uneven number in view
of the fact that uneven numbers are comparative rarities. If both paternal and maternal

216 MONTGOMERY A STUDY OF THE CHROMOSOMES

germ cells gradually acquired the same uneven number of chromosomes, then by conju-
gation of such cells, similar numbers of chromosomes being added together, a new even
number would result. But there is still another possibility by which the uneven number
could pass into an even one. The odd chromosome, at least in the cases here desrrilx !,
does not divide in the second maturation division, and so behaves abnormally. Now such
an abnormally behaving chromosome might in time become differentiated from the other
chromosomes, and I venture the view that such odd chromosomes are on the way to be-
come chromatin nucleoli. The main fact on which this conclusion is based is that in
Protenor belfrageiii is the odd, the eleventh chromosome the "chromosome x" which
in the spermatocytic growth period evinces the phenomena of a chromatin nucleolus.
Then another correspondence is that the chromatin nucleoli in most Hemiptera act like
the odd true chromosome in usually not dividing in the second maturation division. Here
we have an explanation for the origin of those peculiarly modified chromosomes, the chro-
matin nucleoli, thoroughly in accord with the facts I have described for them : the chro-
matin nucleoli are modified chromosomes, in point of origin the odd chromosomes which
appear in the period of transition from a higher (or a lower) to a lower (or higher) even
normal chromosomal number. And there are generally two chromatin nucleoli in the
spermatogonia, because the odd chromosome in those cases where there is an uneven nor-
mal number had probably been formed in most cases, as it certainly appears to have orig-
ignated in Protenor belfragei, as a union of two univalent, chromosomes which had failed
to separate from one another in the spirem stage of the spermatogonic mitosis. This also
explains why the two chromatin nucleoli are generally placed close together in the monas-
ter stage of the sperrnatogonium, they having been originally contiguous in the spirem
thread.

Such an explanation of the origin of the chromatiu nucleoli from the odd chromosomes
seems to be in accord with all the facts, and so far may be considered a true explanation.
The chromatin nucleoli are modified chromosomes, and it is the odd chromosomes which
become thus modified. Conversely, we should expect that chromatin nucleoli would be
formed whenever the chromosomal number in changing from a higher to a lower one, or
the converse, passes through a transition period of an uneven number. Now, as has been
shown in the descriptive part of this paper and tabulated on page 207, all the Hemiptera
sxamined have two chromatin nucleoli, but some have a larger number. Wherever
there is a larger number we find generally that they are of different volumes, and the
question arises : why this difference in volume? The explanation might be that the largest
ones are those most recently formed ; the smallest those which had been evolved at earlier
periods, and which are smaller because they are perhaps diminishing through a gradual
degeneration. If the chromatin nucleoli when once formed should always preserve their

OF THE GKEM CELLS OF METAZOA. 217

original size, there should be no gradations in volume no degeneration on their part so
that in a given species we could determine by their number how many times the chromo-
somal number had changed. But when new chromatin nuo.leoli are formed, the older ones
would seen} to degenerate in the order of their formation. This assumption would explain
the occurrence of very minute chromatin nucleoli found in cells of certain Hemiplcra along
with much larger ones ; the minute ones would represent chromatiu nucleoli formed at
earlier periods, now on the way to total degeneration and disappearance. We might ex-
plain the general occurrence of one pair in the spermatogonia, or of one bivalent one in
the spermatocytes, on the conclusion already reached in an earlier part of this paper, that
the chromatin nucleoli are metamorphosed for a special function different from that of the
other chromosomes and so necessary for the nuclear activity ; and the reason for their de-
generation when new ones are formed, in that a single pair would generally appear to be
sufficient for this function, so that not more than one pair would remain in functional
activity at one time.

Thus we find that the unexpected discovery of an uneven chromosomal number in
the spermatogonia opens the way to an explanation of certain phenomena, and suggests
others not anticipated. It suggests that there is a gradual evolution in the numbers of
chromosomes ; that they have not been fixed from the start, but that with the evolution
of the species the chromosomal number changes and at each change probably passes
through a period with an uneven number. In the Hemiptera this would seem to be, in
the forms examined, a change from higher to lower numbers, and in such a change the
odd chromosome becomes metamorphosed becomes a metamorphosed chromatin nucleolus.
If attention be given to these points in other groups of animals, there can be little doubt
that there, too, will be found occasional examples of uneven normal chromosomal num-
bers, and probably also in some of these cases the production of structures comparable to
the chromatin nucleoli of the Insects. There is great need, first of all, however, to deter-
mine for the Hemiplera whether in such cases there is a close correspondence between
the spermatogonesis and ovogenesis that correspondence I have assumed, since I have not
studied the ovogenesis.

4. Considerations on the Cycle of the Germ Cells.

Here shall be considered in succession some points of broader interest which have
arisen in the course of my studies on spermatogenesis.

(a) The sequence of the stages of the cycle.

In the germ cells of the Metazoa there may be seen regular cycles of generations
following upon one another. In each cycle may be noted a stage of conjugation of ma-

218 MONTGOMERY A STUDY OF THE CHROMOSOMES

ternal and paternal cells or stage of fertilization ; upon this follow a number of ovogonic
or spermatogonic generations, the exact number of which has not yet been determined for
any rnetazoon ; the last generation of the ovogonia or spermatogonia give rise to ovocytes
or spermatocytes of the first order, and these are characterized by the synapsis stage and
growth period when the reduction in the number of chromosomes is effected, the synapsis
stage being evidently coincident in all forms with the commencement of the growth
period ; and finally occur two maturation divisions which result in the formation of ovo-
tids or spermatids. The spermatids undergo an elaborate metamorphosis to become sper-
matozoa ; but since such a metamorphosis is not found in the ovotids, we may disregard
this stage, which evidently is far less conservative than the others ; from the comparative
standpoint the metamorphosis of the spermatozoon is of much less morphological signifi-
cance than the preceding stages of spermatogenesis, and would appear from the recent
investigations to be far more variable.

Thus each germinal cycle shows the following well-marked stages: conjugation or
fertilization, a stage of a number of ovogonic or spermatogonic generations, the synapsis
stage coincident with the growth period, and the stage of the two maturation divisions.
Each such cycle is succeeded by a similar one, and so on indefinitely for an indefinite
number of cycles. Now it is unthinkable that a cycle should be without a beginning ; it
must have been gradually evolved, and some particular stage in it must have been the
starting point. What was this first stage? An answer is necessary before we can enter
into the discussion of the meaning of the synapsis stage.

It appears to me most probable that the stage of conjugation of the germ cells must
be considered the starting point. For from the studies of R. Hertwig and Maupas
on Infusoria, it appears probable that conjugation or fertilization is essentially a pro-
cess of rejuvenation : cells may divide and reproduce for a number of generations
asexually, but there comes a period when the cellular vitality diminishes, so that no further
reproduction is possible except after rejuvenation afforded by conjugation with another
cell. When thus rejuvenated by admixture of substances from the other conjoint, the
cell starts upon a new period of generation the period of conjugation thus being the
commencement of a cycle. As we shall see, the synapsis stage is really a delayed part of
the process of conjugation, and the growth period is induced by the synapsis of the chro-
mosomes. Having determined the starting point of the germinal cycle, we may now con-
sider the meaning of the synapsis stage.

(b) The phylogeny of chromosomes and the significance of the synapsis stage.

In the considerations that follow I assume that through the germinal cycle the chro-
mosomes preserve their individuality from generation to generation i. e., that a particular

OF THE GERM CELLS OF METAZOA. 219

chromosome of one generation is represented in a particular one of a preceding, so that
chromosomes are not produced de novo in each generation. The evidence for this assump-
tion, as regards the Hemiptera, has heen already stated above (cf. the heading : " The
process of sperm atogenesis in the Hemiptera ") ; other evidence was shown in my study
on Peripatus (1901), and there also the observations of other workers was considered in
some detail so it is not necessary at this point to enter into these particulars. Without this
assumption, which is an actuality, as I have shown in some cases, it would be very diffi-
cult to determine the meaning of the stages of the germinal cycle ; while on this assump-
tion much becomes clear, and the phenomena of the synapsis stage alone are strongly
corroborative of this assumption.

Now in the cycle of the germ cells there is a chromosomal peculiarity which has been
described by other investigators, but its significance has not been understood ; I referred
to it in my study of Peripatus (1901). In the anaphases of the male and female pro-
nuclei, as in the anaphases of the early cleavage cells, it is characteristic that each chro-
mosome becames vesicular so that at this stage each daughter nucleus appears composed
of as many such vesicles as there are chromosomes. Each vesicle has its own limiting
wall, and not infrequently the different vesicles may be only loosely connected together ;
ultimately, however, when the complete rest stage is attained, the boundaries between the
vesicles disappear so that the nucleus appears a whole without separated parts.*

Riickert (1895) supposes the chromosomal vesicles to represent a shortened anaphase,
occasioned by the rapid sequence of the mitoses in the blastonieres ; that this is hardly a
correct explanation is seen from the following considerations. From the list of cases just
mentioned in the footnote it will be seen that anaphases with vesicular chromosomes are
found in the pronuclei and in the earlier cleavage cells i. e., in nuclei at the beginning of
the germinal cycle. I have never seen such vesicular stages in the last generations of
ovogonia and spermatogonia, nor to my knowledge has any one else ; but in these later

* This vesicular stage of the chromosomes in the anaphases of mitosis has been described by the following
workers, though this is probably not a complete list : Remak (1855, cited by Heuneguy, 1896, blastomeresofBafracAta);
Oellacher (1872, egg of Trout); <Trinchese (1875, cited by Ilenneguy. 1896, pole cells of Aeolididae); O. Hertwig
(1876, id. citat., blastonieres of Bufo}; Fol (1879, id. citat., Toxopneustei egg); Henneguy (1882, 1891, egg of Trout);
Bellonci (1884, cited by Heuneguy, 1896, blastomeres of Axolotl); Schwarz (1888, -id. citat., blastomeres of Trout);
Van der Stricht (id. citat., larval epidermic of Salamandra and Triton, megacaryocy tes, leucoblasts and erythroblasts
of embryonic liver of Mammals); Mead (1895, 1898, Chaitopterus, female prouucleus and blastonieres up to 16-cell
stage); Foot (1894, 1897, female pronucleus of Allolobopfiora); Sobotta (1897, pronuclei and first cleavage of AmpJii-
oxus); Kostanecki and Wier/ejski (1896, male and female pronuclei of Physa); v. Klinckowstroui (1897, male and
female \>Toanc\eiof Prosthoceraus); Riickert (1895, blastomeres of Cyclops); O. Schultze (1887, blastomeres of Axolotl);
Koelliker (1889, blastomeres of Siredon); Van Beneden and Neyt (1887, blastomeres of Ascaris); Bourn (1888, male
and female pronuclei and first cleavage of Petromyzon); Wheeler (1897, female pronucleus of Myzostoma, occasion-
ally showing widely separated chromosomal vesicles) ; Coe (1898, female pronucleus and first cleavage of Cerebratului) ;
Boveri (1888, blastomeres of Atcarit),

220 MONTGOMERY A STUDY OF THE CHROMOSOMES

germinal stages, as well as in apparently all adult tissue cells (compare Flemming, 1882,
1890; Zimmermann, 1890; Rabl, 1885 ; Torok, 1888), there is generally no such vesicular
stage during the anaphase it is then characteristic of embryonal cells, of those at the
commencement of the generative cycle. An explanation of a possible reason for the
chromosomal vesicles maintaining their independence was given in my paper on Pervpafau
(1901), where I referred it to the breaking of the linin spirem effected by the reduction
mitosis, and maintained that no continuous chromatin spirem could be formed i. e., no
close juxtaposition of the chromosomes be effected until the linin spirem had become
restored.

Probably the chromosomal individuality is maintained through all the generations of
the cycle, but the chromosomes seem to show their independence most markedly in the
early stages, where it is strikingly evinced by their vesicular phenomena. Each vesicle
appears to be potentially a little nucleus, with its ova wall, its chromatic reticulum and
caryolymph, and sometimes with its own nucleoli. This is very suggestive of the possi-
bility that each chromosome may represent, from the phyletic point of view, a nucleus ;
and a metazoan nucleus would then be a symbiotic union of as many nuclei as there are
chromosomes. Such a conclusion might explain why the chromosomes pass through
vesicular phases resembling nuclei in the earlier periods of the cycle.

So far we have seen that in the earlier portion of the germinal cycle the chromosomes
remain more disconnected from one another than at later periods ; in the later periods,
those e. g. of the last generations of the spermatogonia and ovogonia, they no longer show
vesicular, nuclear-like appearances in the anaphases, and appear to be more dependent
upon one another less independent. Now another line of facts may be considered in
this regard. Van Beneden (1883, 1887) first showed that in the fertilized egg of Ascaris
the paternal and maternal chromosomes remain separated from one another, so that in the
prophases of the first cleavage mitosis a paternal and a maternal chromatin spirem is
formed ; thus Van Beneden concluded a maintenance of the individuality of the pro-
nuclei. Then Ruckert (1895) found in the cleavage cells of Cyclops that the paternal
and maternal chromosomes form two separate groups throughout the mitosis, and that
even in the rest stage there is a double nucleus, half paternal and half maternal ; in the
prophases there is a paternal chromatin spirem distinct from the maternal one. Up to
about the 32-cell stage Riickert was able to find these double nuclei, but found that in
later cleavage stages they gradually decrease in number. But Ruckert is probably in
error when he concludes that the separation of the paternal and maternal chromosomes is
retained even up to the time of the first maturation mitosis (first pole spindle). He bases
this conclusion on the discovery that in the equatorial plane of the spindle at this stage
the chromosomes are arranged "ausnahmslos" into two groups. Now here the cbromo-

OF THE GERM CELLS OF METAZOA. 221

somes are bivalent and eleven in number, so that each group cannot have the same num-
ber. Thus out of the cases examined by him, in twelve cases the chromosomes were in
groups of relatively equal number (relation of six to five) ; in two cases, one group had
four, the other group seven chromosomes ; in three cases, they were arranged in groups of
three and eight, respectively ; and in one case, in groups of two and nine, respectively.
Thus, though there may be at the period of the first maturation mitosis an arrangement
of the chromosomes into two groups, yet the variable discrepancy in the number of the
chromosomes composing the two groups shows that it is impossible that one group has only
paternal chromosomes and the others only maternal, for the reason that at the start (in
the fertilized egg) paternal and maternal chromosomes are equal in number. Accordingly,
Riickert has shown for Oyclops that the maternal and paternal chromosomes form separate
groups up to about the 32-cell stage, when the separateness of these groups gradually
disappears ; and his own descriptions and figures would show that at the time of the
first maturation mitosis there is no longer a paternal group of chromosomes separate from
a maternal group. There could also be mentioned the observations of other authors to
the effect that the paternal and maternal chromosomes compose separate groups in the
early cleavage cells, as especially the observations of my colleague, Prof. Conklin, on the
eggs of Crepidula.

Accordingly, we have seen that in the earlier period of the germinal cycle, at the
time of fertilization and the immediately following generations, paternal and maternal
chromosomes remain separated from one another, and also that the individual chromo-
somes show a remarkable degree of independence as evinced by their vesicular phenomena
in the anaphases. In the later stages of the germinal cycle, on the contrary, paternal and
maternal chromosomes appear no longer to be arranged in separate groups, and the chro-
mosomes themselves are no longer vesicular in the anaphases.

Now for the bearing of all this on the question of the significance of the synapsis
stage. At the commencement of the germinal cycle, the stage of conjugation of the germ
cells, the chromosomes are more distinct from one another than at any later stage ; this
distinctness gradually disappears as the cycle progresses, and at the time of the synapsis
stage the chromosomes actually join together to form half the normal number of (bivalent)
chromosomes. What chromosomes are these which unite to form pairs ? Does a paternal
chromosome unite with a paternal and a maternal with a maternal, or does a paternal chro-
mosome unite with a maternal one ? The following considerations show that the latter
view is probably the true one.

First of all, in Ascaris megaloc&phala univalens there is the normal number of two
chromosomes ; as Brauer (1893 b) has demonstrated, one of these is paternal, the other
maternal in origin ; since these two unite to form one bivalent one in the synapsis stage,

222 MONTGOMERY A STUDY OF THE CHROMOSOMES

this would be a union of a paternal with a maternal chromosome. Also in the Hemiptera,
whenever there are in the spermatogonia two chromosomes which are distinguishable from
the others by their greater size, as in several species described in this paper e. g., Protenor
belfragei these two especially large ones always unite together in the synapsis to form
one bivalent one much larger than the other bivalent ones, and one ot the large ones does
not unite with a small one ; now it can be shown that one of these large chromosomes is
paternal and the other is maternal. For calling the two large chromosomes of the sper-
matogonia a and b, respectively, they unite in the synapsis to form the bivalent chromo-
some ab ; the first maturation mitosis (here a reduction division) gives a to one daughter
cell (second spermatocyte) and b to the other ; the second maturation division of the one
of these daughter cells gives to each spermatid ia, the corresponding division of the other
daughter cell hb to each spermatid. What AVC find in each of these spermatids is only
one especially large chromosome, and not two. Accordingly, in order for there to be two
in the spermatogonia, the egg cell must furnish one, and then that one, together with the
one furnished by the spermatozoon in fertilization, would make up the two. Then of the
two particularly large chromosomes of the spermatogonium, one would be paternal and
one maternal, and since these two unite in the synapsis stage, this would be a union of a
paternal chromosome with a maternal chromosome. A case where two particularly large
chromosomes are distinguishable in the spermatogonia was selected for discussion, because
these two, on account of their peculiarity in size, can be recognized through the matura-
tion divisions ; but if the conclusion be true that one of the large chromosomes is paternal
and one maternal, and that these two join together in the synapsis, then it would be very
probable that each of the other bivalent chromosomes of the spermatocytes represents a
univalent paternal chromosome united with a univalent maternal one. This case, as the
one of Ascaris megalocephala univalens, may be considered very positive cases in favor of
the union of paternal with maternal chromosomes in the synapsis stage. There is still
another point of view which makes this conclusion very probable. As we have seen,
whenever there is an even number of chromosomes in the spermatogonia, exactly half that
number of bivalent chromosomes are formed in the synapsis ; thus in Euchistus variolarius
there are fourteen univalent chromosomes in the spermatogonia, and seven bivalent ones
in the first spermatocytes. Now seven of these chromosomes are paternal and seven ma-
ternal, since the spermatids have only seven. The regular formation of seven bivalent
chromosomes in the synapsis stage would be only possible if maternal chromosomes united
with paternal ones. For if, on the contrary, paternal chromosomes united with paternal
and maternal with maternal, then of the seven paternal chromosomes three bivalent ones
could be formed, but there would be left an ununited odd one, and similarly of the seven
maternal chromosomes there would remain an ununited (univalent) odd one. But since

OP THE GERM CELLS OF METAZOA. 223

in the first spermatocytes of Euchistus all the chromosomes are united into pairs, so that
there are no chromosomes remaining univalent, it follows that in this case it is impossible
that only chromosomes of like parentage should unite together.

These considerations render it very probable that in the synapsis stage is effected a
union of paternal with maternal chromosomes, so that each bivalent chromosome would
consist of one univalent paternal chromosome and one univalent maternal chromosome.
This conclusion allows us to consider the synapsis stage in an entirely new light, and gives
an important significance to this stage. The synapsis stage then, which is characterized
by the union of chromosomes into bivalent pairs, may be considered the stage of the con-
jugation of the chromosomes. When the spermatozoon conjugates with the ovum there is
a mixture of cytoplasm with cytoplasm, of karyolymph with karyolymph, possibly also
an intermixture of other substances ; but there is then no intermixture of chromatin, for
the chromosomes then, as we have seen, remain more separated from one another than at
any other stage in fact, the paternal chromosomes seem to show a repulsion for the
maternal, inasmuch as they are arranged in two separate groups. But after this begin-
ning stage of the germinal cycle, the repulsion of the paternal for the maternal chromo-
somes gradually diminishes, is generally no longer recognizable in the last of the spermato-
gonic and ovogonic divisions, and in the synapsis stage instead of a repulsion we find a posi-
tive attraction between the paternal and maternal chromosomes. The reason for the final
union of these chromosomes is obvious : it is evidently to produce a rejuvenation of the
chromosomes. From this standpoint the conjugation of the chromosomes in the synapsis
stage may considered the final step in the process of conjugation of the germ cells. It is
a process that effects the rejuvenation of the chromosomes ; such rejuvenation could not
be produced unless chromosomes of different parentage joined together, and there would
be no apparent reason for chromosomes of like parentage to unite. At the same time the
so-called " reduction in the number " of the chromosomes is effected, but this is probably
not primal but rather a necessary result of the conjugation of the chromosomes. And
here the point may be made that really there takes place no reduction in the number of
the chromosomes in the germinal cycle, but " reduction in number " is simply a conveni-
ent phrase for expressing that in the synapsis the chromosomes unite to form pairs ; no
chromosomes have been lost, there is in the strict sense no .reduction in number.

So we find that the synapsis stage has a very broad and important significance, of all
the stages in the germinal cycle second only to the initial stage of conjugation of the germ
cells. In the synapsis stage we see the final process in the conjugation of the germ cells,
namely, the conjugation of the chromosomes. Now following immediately upon the syn-
apsis stage comes the growth period of the spermatocytes and ovocytes that period when
the germ cells attain volumes greater than at any other period in the germinal cycle.

224 MONTGOMERY A STUDY OP THE CHROMOSOMES

Very evidently this great increase in volume is effected by that rejuvenescence of the chro-
mosomes secured by their conjugation. For the chromosomes are centres of metabolic
activity, by conjugation of paternal with maternal chromosomes in the synapsis stage
their metabolic functions are rejuvenated, and this rejuvenation finds its expression in the
great changes of the growth period. So this explanation of the. synapsis stage would seem
to be in accord with all the facts known at present.

It is quite possible that at an earlier period in the phylogeny, the conjugation of the
chromosomes may have taken place at the time of the conjugation of the germ cells, and
not have been separated from that stage by a number of generations as in the modern
Metazoa. But the determination of the original time of occurrence of the conjugation of
the chromosomes is highly speculative, and so will not be entered upon here.

It is generally stated (e.g. Von Rath, 1893 ; Riickert, 1894) that the bivalent chro-
mosomes of the spermatocytes and oyocytes of the first order are produced " by the spirem
segmenting into only half the normal number of chromosomes." This is not a correct
statement, since in the prophases of the first maturation mitosis there is, as I have shown
in my paper on Pcripatus, no stage of a continuous chromatin spirem. Further, this
general statement is not at all explanatory of the formation of bivalent chromosomes, for
it does not express any reason why the chromosomes should be joined into pairs.

It is to Moore (1895) that we owe the first clear characterization and estimation of
the synapsis stage ; he divided the germinal cycle into the " first period " (conjugation of
germ cells, spermatogonic and ovogonic divisions), the "synaptic phases" (coincident with
the growth period), and the " second period " (maturation divisions). It will be seen that
my own classification of the stages is somewhat more detailed than Moore's, though it is
not necessarily any better. The important characteristic of the synapsis stage is, of course,
the union of chromosomes into bivalent pairs; the exact details of this process, which
appear to differ in different groups, are of secondary significance.

(c) The significance of the maturation divisions.

The two maturation divisions of the Metazoa represent the terminal stages of the
germinal cycle.

In the Copepods (Hacker, 1895; Eiickert, 1894), the Isopods (Oniscus, in a just
finished paper by my student, Miss Louise Nichols), in the Insects (Von Rath, 1892 ;'
Henking, 1890; Montgomery, Paulmier, 1899 ; McClung, 1900), and in Peripatus (Mont-
gomery, 1901) there are well demonstrated cases that one of the maturation divisions is a
reduction division (pseudoreditclion, Riickert, 1894) in that it accomplishes a separation
of entire univalent chromosomes from one another. Such a reduction division, a trans-
verse splitting of the chromosomes, is not known for any other generation of the germinal
cycle, nor for any somatic generation.

OF THE GEKM CELLS OP METAZOA. 225

But in Ascaris (Braucr, 1893 b), in Salamandra (Meves, 1896), in the Rat (Len-
hossek, 1898), in Selach'd (Moore, 1895), and in Amphiuma (McGregor, 1899) the au-
thorities cited agree that both maturation divisions are equational. Now it does not seem
a priori probable that in some Metazoa a reduction division should occur, and in others
not. The case of Ascaris would seem to show no sign of a reduction division, for Brauer's
careful study apparently shows that each bivalent chromosome becomes split longitudi-
nally twice ; yet Sabaschnikoff 's more recent study (1898) would show that another inter-
pretation of the phenomena is possible (but not proved), namely, that the chromatin
microsomes may become rearranged into fours in such a way that one of the maturation
divisions may be reductional. In the Salamander, Flemming (1887) showed that the
mitosis of the first generation of spermatocytes has remarkable peculiarities, so that he
named it a " heterotypic " mitosis. The most remarkable of its characteristics is that the
chromosomes are longitudinally split in shape like horseshoes, and that they open up into
the forms of rings, the ends of the daughter horseshoes retaining their mutual connection.
Such a heterotypic mitosis was corroborated by Meves in his description of the spermato-
genesis of Salamandra ; and it has been shown to be characteristic of the first maturation
division in Selachii, the Rat, and Amphiuma by Moore, Lenhossek and McGregor, respec-
tively.* All these writers show that the heterotypic mitosis results in a longitudinal
division of the chromosome, and there can be no reasonable doubt of the correctness of
their descriptions. But I think they are mistaken in concluding that because the hetero-
typic division is a longitudinal division of the chromosomes, that therefore it is an equa-
tion division. For the chromosomes of these spermatocytes are bivalent there are just
half as many in the spermatocytes as in the sperruatogonia. Since there is no loss of
chromosomes in the spermatocytes, there must take place a union of univalent chromo-
somes into pairs during some part of the growth period i. e., in the synapsis stage. I
venture the view that in the Vertebrates either (1) the bivalent chromosomes are formed
by every two univalent chromosomes becoming apposed to one another side to side i. e.,
along their whole length, so that the two would compose a double horseshoe or (2) by the
two ends of one univalent chromosome becoming closely connected with the two ends of the
other, so that the whole would have the form of a ring. From what has been described
for these bivalent chromosomes, we know that the longitudinal split does not divide their
ends, but the ends are unsplit. Accordingly, it would appear probable that the essential
process in the formation of these bivalent chromosomes is that the two ends of one univa-
lent chromosome become united with the two ends of another, while it would be of secon-
dary importance whether the two chromosomes might be apposed along their whole lengths
or not.

* Moore states that the second maturation division is also heterotypic, but his figures do not prove his point,
which needs ree'xamination.

226 MONTGOMERY A STUDY OF THE CHROMOSOMES

But from this it would follow that the heterotypic mitosis of the first spertnatocytes
of the Vertebrates is really a reduction division, and results in the separation of whole
univalent chromosomes. Then the longitudinal split of such bivalent chromosomes would
be really the space between two univalent chromosomes. Thus, though these chromosomes
may appear in the prophases longitudinally split, yet a separation of the daughter chromo-
somes along the line of this split would not be an equational division. The workers on
vertebrate spermatogenesis have indeed shown that the bivalent chromosomes are split
longitudinally, but since none of them have succeeded in demonstrating h<>\v the bivalent
chromosomes are formed in the synapsis stage they could not show the ni^iiil'icance of
this longitudinal split. For Peripatus and the Hemiptera I have shown that a bivalent
chromosome is produced by one end of one univalent chromosome uniting with one end
of another ; while in the Vertebrata, if my interpretation is correct, a bivalent chromosome
would be produced by the union of both ends of one univalent chromosome with both
ends of another the spermatogonic chromosome is U-shaped, the spermatocytic chromo-
some is ring-shaped since it represents two such U-shaped elements with their ends coii-
nected. Also in Peripatus and the Hemiptera there are occasionally ring-shaped chro-
mosomes similar to the heterotypic chromosomes of Vertebrates, and they are formed by
the two ends of one univalent chromosome being joined with the two ends of another,
instead of one end being joined simply with one end.

This interpretation explains, and the process has never been satisfactorily explained
before, why one of the maturation mitoses in Vertebrates is heterotypic : it is a reduction
division separating entire univalent chromosomes, and it differs from all other mitoses
of the germinal cycle because it is the only one of them which does separate entire chro-
mosomes. If this view is true, then probably all Metazoa would have in common the
occurrence of one reduction division, and we should no longer be confronted by the dis-
crepancy between Metazoa with and those without a reduction division. The occurrence
of a reduction division is actually proved for the Copepoda, Insecta, Oniseus and Peri-
patus (I shall not mention other objects where it has been rendered probable but not
thoroughly proven) ; a priori we should expect that one would occur in the Vertebrates
also, and in the Vertebrates there does occur a peculiar heterotypic division which, as
we have seen, can be satisfactorily explained as a reduction division. Accordingly, the
term " heterotypic mitosis" might be applied to any mitosis which results in the separation
of whole univalent chromosomes, irrespective whether it divides the bivalent chromosome
transversely or longitudinally ; the term " heterotypic " is indeed most excellent in that
it expresses a mitosis " of a type of a different kind," one differing from all other mitoses
of the germinal cycle. Of very secondary importance, then, would be the form of the
chromosomes not the form but the way in which the chromosomes divide should be

OF THE GERM CELLS OF METAZOA. 227

taken as the criterion of the heterotypic mitosis. This signification would be different
from that originally defined by Flemming (1887), but it would certainly be a step toward
greater clearness to use " heterotypic " division in the place of the promiscuously used
" reduction " division.

Now that we have seen that a reduction division occurs in the Cop&poda, Insecla, Onis-
cus and Peripatus, and that the heterotypic division of the Verlebrata may be interpreted
as a reduction division also, we have to try to explain why such a reduction division occurs.
In the synapsis stage there is a conjugation of paternal with maternal chromosomes for the
purpose of rejuvenation of the chromosomes as metabolic centres, and this rejuvenation is
exemplified in the great metabolic activity of the growth period. Now, It. Hertwig and
Maupas have shown for the Infusoria that the two conjoiuts remain for only a certain
period in apposition, and that when the interchange of nuclei necessary for rejuvenation
has been accomplished the conjoints separate. Of course it is not a true analogy to com-
pare conjugating Infusoria (i. e., whole cells) with conjugating chromosomes (i. e., por-
tions of cells). But still it is very probable that two chromosomes unite temporarily for
the same reason that two Infusoria do, that is, for an interchange of substances ; and
when the chromosomes have accomplished this interchange there would no longer be
any necessity for continued apposition, so they tend to separate from one another. It is the
reduction division in Metazoa which accomplishes the complete separation, though it may
commence in the prophases of this division. It is conceivable that the conjugated chro-
mosomes might separate as they had come together, without the intervention of a mitosis.
But in the Metazoa, so far as we know, they become separated only by the agency of a
mitosis, and that is the reduction mitosis. At the beginning of the germinal cycle there
is a repulsion of paternal and maternal chromosomes for one another, during the synapsis
a strong attraction, and at the end of the germinal cycle a repulsion again, but not a re-
pulsion so strong as to distribute the chromosomes into two groups in the spermatid and
ovotid. Only by a reduction division can paternal and maternal chromosomes become
wholly separated, for only then do the interchromosomal linin fibres (persisting portions

f

of the linin spirem : compare my paper on Peripatus) become broken.

The question is complicated because another maturation division occurs, an equa-
tional division : in the Insecta, Oniscus, Peripatus and the Vertebrates the reduction
division precedes the equational, in the Copepoda (according to Riickert, 1894) the re-
verse is the case. It is not difficult to explain why an equation division should occur at
this time, for the cell has increased in volume very greatly during the growth period, and
great increase in volume (increase beyond the individual mass) would appear to be a main
factor in inducing cell division. With this increase in volume of the cell the chromo-
somes also increase in volume (though by no means in a direct ratio), and each univalent

228 MONTGOMERY A STUDY OF THE CHROMOSOMES

component of a bivalent chromosome divides into two longitudinally (equationally), this
being the usual mode of division of a chromosome in fact the only known method of re-
production of univalent chromosomes. From this standpoint the growth period would
be the inducer of the equational maturation mitosis, and this mitosis would be strictly
comparable physiologically to any other equation mitosis of the germinal cycle. But at
the time that the cell is preparing for this equation division, paternal and maternal
chromosomes, having accomplished (he purpose which occasioned their conjugation, show
a tendency to repulsion for each other, and so evince the need of becoming disconnected.
The cell already started into mitotic activity would offer mechanical possibilities to effect
the separation of the maternal from the paternal chromosomes, so that instead of a single
mitotic process there are two in rapid succession, sometimes with not a trace of a rest stage
in between, one separating entire univalent chromosomes, the other separating the halves
of each univalent chromosome. It would be very enticing to enter here into the me-
chanics of this mitosis, which would be practically a determination of the points of appo-
sition of the mantle fibres on the chromosomes ; but such an inquiry is hardly germane
to the present discussion. The point to be made is that in the Metazoa there follows
after the growth period an equation mitosis, because in the growth period the cell has in-
creased beyond the normal size ; and that a reduction division occurs about the same time
in order to affect a complete separation from one another of the paternal and maternal
chromosomes, which, having accomplished the purpose for which they conjugated, show
again a repulsion for each other. The growth period is the inducer of the equation di-
vision, the mutual repulsion of chromosomes of different parentage the inducer of the
reduction division. The chromosomes in the late anaphases, after the maturation divis-
ions, become vesicular and so show a great degree of mutual independence, because the
reduction division had severed their linin connections.

In conclusion, it would be very interesting to enter into the question of the parallel-
ism of the germinal cycle in Metazoa and Protozoa, as has been done by Henking and
Moore (1895). However, it would be well first to have ascertained the significance in
the cycle of the Metazoa, as far as that can be done without reference to the states in the
Protozoa. The chromosomes are the cell components on which the problem can be best
studied.

OF THE GERM CELLS OF METAZOA. 229

IV. LITERATURE REFERENCES.

1884. BELLONCI : La Karyokinese dans la segmentation de 1'oeuf de 1'Axolotl. Arcli. ital. de Biol. 6. Accad. del

Lincei.
1888. BOEHM, A. A. : Ueber Reifung und Befruchtung des Eies vom Petromyzon Planeri. Arch. mikr. Anat. 32.

1888. BOVEEI, T. : Zellen Studien, 2. Jena.

1809. - - Die Entwickelung vom Ascaris megalocephala mit besonderer Rucksicht auf die KernverhiiHnisse.

Festschrift von Kupffer, Jena.
1893a. BRAUEK, A. : Zur Kenntniss der Reifung des parthenogenetisch sich entwickelnden Eies voni Artemia salina

Arch. mikr. Anat. 43.

18931). - Zur Kenntniss der Spcnnatogenese vom Ascaris megalocepliala. Ibid., 42.
1898. COE, TV. R. : The Maturation and Fertilization of the Egg of Cerebratulus. Zool. Jahrb. 12.
1882. FLEMMING, W. : Zellsubstanz, Kern und Zelltheilung. Leipzig.

1887. Neue Beitrage zur Kenntniss der Zelle. Arch. mikr. Anat. 29.

1890. Ueber die Theilung vom Pigmentzellen uud Capillanvandzellen. Ibid., 35.

1879. FOL, H. : Recherches sur la fe'condation etc. chez divers anirnaux. Mem. soc. phys. et hist. nat. Geneve, 26.

1894. FOOT, KATHARINE : Preliminary note on the Maturation and Fertilization of the Egg of Allolobophora fce-

tida. Journ. Morph. 9.

1897. - The Origin of'the Cleavage Centrosomes. Hid., 12.

1895. HACKER, V. : Die Vorstadicn der Eireifung. Arch. mikr. Anat. 45.

1894. HEIDENHAIN, M. : Neue Untersuchungen fiber die Centralkorper undihre Beziehungen zum Kern und Zellen-

protoplasma. Arch. mikr. Anat.
1890. HENKING, H. : Untersuchungen fiber die ersteu Entwicklungsvorgange in den Eiernderlnsekten. II. Ueber

Spermatogenese und deren Bcziehung zur EntwicklungbciPyrrhocoris apterus. Zeit. f. wiss. Zool. 51.
1882. HENNEOUY : Sur la formation des feuillcts embryonnaires chez la Truite. C. R. Acad. sci.
1891. Nouvelles recherches sur la division cellulaire indirecte. Journ. Anat. et Phys.

1896. Logons sur la Cellule. Paris.

1876. HERTWIG, O. : Beitrage zur Kenntniss der Bildnng, Befruchtung und Theilung des thierischen Eies. Morph.

Jahrb.
1890. Vergleich der Ei- und Samenbildung bei Nematoden. Arch. mikr. Anat. 36.

1889. HERTWIG, R. : Ueber die Conjugation der Infusorien. Abhandl. bayer. Akad. Wiss. II. Cl. 17.

1897. v. KLINCKOWSTROEM : Beitrage zur Kenntniss der Eireifung und Befruchtung bei Prostheceraeus vittalus.

Arch. mikr. Anat. 1897.

1889. KOELLIKER, A. v. : Handbuch der Gewcbelehre des Menscheu. 6tc Auflage.

1896. KOSTANECKI and WIERZEJSKI : Ueber das Verhalten der sogen. achromatischen Substanz im befruchteten Ei.
Arch. mikr. Anat. 47.

1898. LENHOSSEK, M. v. : Untersuchungen uber Spermatogenese. Ibid., 51.

1889. MAUPAS, E. : La Rajeunissement karyogamique chez les Cilies. Arch. Zool. Exp6r. (2) 7.

1899. McCLUNG, C. E. : A peculiar Nuclear Element in the male reproductive cells of Insects. Zoiilog. Bulletin.

1900. - The Spermatocyte Divisions of the Acrididoe. Bull. Univ. of Kansas, Kansas Univ. Quarterly, 9, 1.
1899. McGiiEGOR, J. H. : The Spermatogensis of Amphiurna. Journ. Morph. 15, Supplement.

1895. MEAD, A. D. : Some Observations on Maturation and Fecundation in Chsetopterus pergamentaceus Cuvier.

Ibid., 10.

1898. Origin and Behavior of the Centrosomes in the Annelid Egg. Ibid., 14.

1896. MEVES, F. : Ueber die Entwicklung der mannlichen Geschlechtszellen vom Salamandra maculosa, Arch.

mikr. Anat. 48.
1898a. MONTGOMERY : The Spermatogenesis in Pentatoma up to the Formation of the Spermatid. Zool. Jahrb. 12.

1898b. Observations on various Nucleolar Structures of the Cell. Biol. Lectures Woods Holl. Lab.

1899a. Chroniatin Reduction in the Hemiptera : a Correction. Zool. Anz. 22.

260 MONTGOMERY A STUDY OF THE CHROMOSOMES

1899b. Cytological Studies, with especial Regard to the Morphology of the Nucleolus. Journ. Morph. 15.

1901. The Spermatogensis of Peripatus (Peripatopsis) balfouri up to the Formation of the Spermatid. Zool.

Jahrb.
1895. MOORE, J. E. S. : On the Structural Changes in the Reproductive Cells during the Spermatogenesis of Elas-

mobranchs. Quart. Journ. micr. Sci. (N. S.) 38.

1872. OELLACHER, J. : Beitriige zur Qeschichte des Keimblaschens im Wirbelthiereie. Arch. mikr. Anat. 8.
1898. PAULMIER, F. C. : Chromatin Reduction in the Hemiptera. Anat. Anz. 14.

1899. The Sperraatogenesis of Anasa tristis. Journ. Morph. 15, supplement.

1885. RABL : Ueber Zelltheilung. Morph. Jahrb. 10.

1855. REMAK : Untersuchungen iiber die Entwicklung der Wirbclthiere. Berlin.

1894. RUCKEBT : Zur Eireifung bei Copepoden. Anat. Hefte, IV, 2.

1895. TJeber das Selbstaudigblciben der vaterlichen und mutterlichen Kernsubstanz wiihrend der ereten Ent-
wicklung des befruchteten Cyclops-Eies. Arch. mikr. Anat. 45.

1897. SABASCHNIKOFF, M. : Beitrage zur Kenntniss der Chromatin reduction in der Ovogenese vom Ascaris mega-
locepbala bivalens. Bull. Soc. impe'r. Nat. Moscow.

1887. SCHULTZE, O. : Ueber die Karyokinese in den crsten Zellen des Axolotl. Sitzber. phys. rned. Ges. Wiirzburg.

1888. SCHWABZ, E. : Ueber embryonale Zelltheilung. Mittheil. embryol. Inst. Wien.

1897. SOBOTTA : Die Reifung und Befruchtung des Eies vom Amphioxus lanceolatus. Arch. mikr. Anat. 50.

1888. TOBOK, L. : Die Theilung der rothen Blntzellen bei Amphibien. Ibid., 32.

1875. TBINCHESE : I priini moment! dell' evoluzione nei Molluschi. Atti R. Accad. Lincei.

1883. VAN BENEDEN, E. : Recherehes sur la maturation de I'oeuf, la fecondation et la division cellulaire. Gand.

1887. VAN BENEDEN and NEYT : Nouvelles recherches, etc. Bull. Acad. Roy. Belg. 14.

1892. VON RATH, O. : Zur Kenutniss der Spermatogenese vom Gryllolalpa vulgaris Latr. Arch. mikr. Anat. 40.

1893. Beitrage zur Kenntniss der Spermatogenese vom Salamandra maculosa. Zeit. wiss. Zool. 57.

1900. WALLACE, LOUISE : The Accessory Chromosome in the Spider. Anat. Anz. 18.

1897. WHEELEB, W. M. : The Maturation, Fecundation and early Cleavage of Myzostoma glabrum Leuckart.

Arch, de Biol. 15.

1890. ZIMMERMANN, K. W. : Ueber die Theilung der Pigmentzellen, etc. Arch. mikr. Anat. 36.

EXPLANATION OF THE PLATES.

All figures have been drawn with the camera lucida at the level of the base of the microscope, with the homo-
geneous immersion objective -f t of Zeiss and ocular 4, tube length 180 mm. In the majority only the Chromatin
nucleoli, chromosomes and true nucleoli together with the outline of the nucleus or cell body has been drawn ; and
in the majority of figures representing lateral views of monaster stages, the mantle fibres, connective fibres, and
centrosomes are the only structures shown beside the chromatin elements. For Eucliistui variolariits, however, the
various structures are represented in detail, and this is also the case with some of the figures of various other species.

The following abbreviations have been employed (for others not here mentioned the descriptive text must be

referred to):

C. Mb., cell membrane.

N., true nucleolus (plasmosome).

N. S, chromatin nucleolus (modified chromosome).

N. Mb., nuclear membrane.

Plate 1.

Euchittus variolariug, Figs. 1-19

Fig. 1. Nucleus of spermatogonium, rest stage.
Figs. 2, 3. Pole views of spermatogonic monasters.

OF THE GEEM CELLS OF METAZOA. 231

Pigs. 4, 5. Lateral views of synapsis stages, only a few of the cbromatin elements shown.

Fig. 6. Nucleus, synapsis, showing two whole bivalent chromosomes.

Fig. 7. One whole bivalent chromosome, end of the synapsis stage.

Figs. 8, 9. Postsynapsis stages, in each three whole bivalent chromosomes shown.

Fig. 10. Telophase of the spermatocyte, showing two whole bivalent chromosomes.

Fig. 11. Late telophase, showing eight bivalent chromosomes.

Fig. 12. Rest stage of spermatocyte.

Figs. 13, 14. Early prophases of first maturation division, each figure showing two whole bivalent chromosomes.

Fig. 15. Two whole bivalent chromosomes, a little later stage than the preceding, showing their liuin connections.

Figs. 10, 17. Later prophases, only in Fig. 16 are all bivalent chromatiu elements shown ; in Fig. 17 the eentrosomc

pairs on the surface of the nucleus.

Fig. 18. Lateral view of the monaster of the first maturation mitosis, all chromatin elements shown.
Fig. 19. Pole view of the same stage.

Suchistus tristigmus, Figs. 20-2G.

Fig. 30. Pole view of spermatogonic monaster.

Figs. 21, 22. Nuclei of first spermatocytes, rest stage.

Figs. 23, 24. Pole views of monaster, first maturation division.

Fig. 25. Lateral view of spindle, first maturation division (showing all chromatin elements).

Fig. 26. Second spermatocyte, chromatin elements not definitely arranged in the equator of the spindle.

Podisus spinosus, Figs. S7-S9.

Fig. 27. Pole view of spermatogonic mouaster.

Fig. 28. Telophase of first spermatocyte, nucleus.

Fig. 29. Pole view of monaster, first maturation mitosis.

Mormidea lugens, figs. 30-33.

Fig. 30. Nucleus of spermatogonium, commencement of prophasc.

Fig. 81. Pole view of spermatogonic monaster.

Fig. 32. Nucleus, postsynapsis stage.

Fig. 33. Pole view of monaster, first maturation mitosis.

Peribalus limbolaris, Figs. 34-37.

Fig. 34. Spermatogonic nucleus, rest stage.

Fig. 35. Pole view of spermatogonic monaster.

Fig. 36. Nucleus of first spermatocyte, rest stage.

Fig. 37. Pole view of monaster, first maturation mitosis.

Cosmopepla carnifex, Figs. 88-41.

Fig. 38. Pole view of spermatogonic monaster.

Fig. 39. Nucleus of first spermatocyte, rest stage.

Fig. 40. Lateral view of the spindle, first maturation mitosis, showing all chromatiu elements.

Fig. 41. Pole view of the same stage.

Nezara hilaris, Figs. 42-45.

Figs. 42, 43. Spermatogonic nuclei, commencement of prophase.

Fig. 44. Pole view of spermatogonic monaster.

Fig. 46. Nucleus of first spermatocyte, early telaphase.

232 MONTGOMERY A STUDY OF THE CHROMOSOMES

Brochymoia sp., Fiyn. 40-49.

Fig. 46. Spermatogonic nucleus, rest stage.
Fig. 47. Pole view of Spermatogonic monaster.

Plate II.

Brochymena sp., Figs. 48, 49.

Fig. 48. Nucleus of first spcrmatocyte, telapbase.

Fig. 49. Pole view of monaster, first maturation mitosis.

Perilhu co/ijluens, Figs. 50-53.

Fig. 50. Spermatogonic nucleus, rest stage.

Fig 51. Spermatogonic monaster, pole view.

Fig. 53. Nucleus of first spermatocyte, rest stage.

Fig. 53. Pole view of monaster, first maturation mitosis.

Camts delius, Figs. 54-63.

Fig. 54 Spermatogonic nucleus, rest stage.

Fig. 55. Spermatogonic monaster, pole view.

Figs. 57, 58. Nuclei of first spermatocytes, rest stage.

Figs. 59, 60. Pole views of monasters, first maturation mitosis.

Fig. 61. Lateral view of the same stage, showing all the chromatin elements.

Fig. 62. Pole view of monaster, first maturation mitosis.

Fig. 63. Nucleus of first spermatocyte, rest stage,

Trichopepla semimttata, Figs. C4-69.

Fig. 64. Spermatogonic nucleus, rest stage.

Fig. 65. Spermatogonic monaster, pole view.

Fig. 66. Nucleus of first spermatocyte, synapsis.

Fig. 67. Nucleus of first spermatocyte, rest stage.

Fig. 68. Pole view of monaster, first maturation mitosis.

Fig. 69. Lateral view of slightly earlier stage, showing all the chromatin elements.

Eurygazter alternatvs, Figs. 70, 71.

Fig. 70. Nucleus of first Bpermatocytr, rest stage.

Fig. 71. Pole view of monaster, first maturation mitosis.

Anata tristis, Figs. 7S-7G.

Figs. 72, 73. Spermatogonic nuclei, rest stage.

Fig. 74. Spermatogonic monaster, pole view.

Fig. 75. Nucleus of first spermatocyte, telaphase

Fig. 76. Pole view of monaster, first maturation mitosis.

Anaia armigera, Figs. 77, 7S.

Fig. 77. Spermatogonic monaster, pole view.

Fig. 78. Pole view of the spindle of the first maturation division, the chromatin elements not yet definitely ar-
ranged in the plane of the equator.

OF THE GERM CELLS OF METAZOA. 233

Anasa sp., Figs. 70-S3.

Fig. 79. Sperrnatogonlo nucleus, rest stage.
Fig. 80. Spermatogouic monaster, pole view.
Fig. 81. Nucleus of first spermatocyte, rest stage.
Fig. 82. Pole view of monaster, first maturation division.

Fig. 83. Lateral view of four bivalent chromosomes, monaster stage of first maturation mitosis^ the poles of the
spindle outside of the plane of the section.

Metapodius tcrminalis, Figs. S4-S7.

Fig. 84. Spermatogouic nucleus, rest stage.

Fig. 85. Spermatogonic monaster, pole view.

Fig. 80. Nucleus of first spermatocyte, telaphase.

Fig. 87. Pole view of monaster, first maturation mitosis.

CJiariesterus antennalor, Figs. SS-90.

Fig. 88. Nucleus of first spermatocyte, telaphase.

Fig. 89. Lateral view of monaster stage of the first maturation mitosis, showing all the chromatiu elements.

Fig. 90. Pole view of the same stage.

Alydus pttosuliif, figs. 91-95.

Fig. 91. Spermatogonic nucleus, rest stage.

Fig. 92. Spermatogonic monaster, pole view.

Figs. 93, 94. Nuclei of first spermatocytcs, rest stage.

Fig. 95. Pole view of monaster, first maturation mitosis.

Plate III.

Alydus eurinus, Figs. 06-102.

Fig. 96. Spermatogonic monaster, pole view.

Fig. 97. Nucleus of first spermatocyte, telaphase.

Fig. 98. Pole view of monaster, first maturation mitosis.

Fig. 99. Lateral view of the same stage.

Fig. 100. Pole view of monaster, second maturation mitosis. ,

Figs. 101, 103. Spermatids at close of second maturation mitosis.

Corizus lateralis, Figs. 103-106.

Fig. 103. Nucleus of first spermatocyte, tclaphase.
Figs. 104, 105. Pole views of monasters, first maturation mitosis.

Fig. 106. Oblique view of the spindle of the first maturation mitosis, before the chromosomes have taken their
definite position in the equator.

Uarmostes reflexulus, Figs. 107-117.

Fig. 107. Spermatogonic nucleus, rest stage.

Figs. 108-110. Spermatogonic monasters, pole views.

Fig. 111. Nucleus of first spermatocyte, rest stage.

Fig. 112. Pole view of monaster, first maturation mitosis.

Fig. 113. Lateral view of the same stage, showing all the chromatin elements.

234 MONTGOMERY A STUDY OF THE CHK( >Mos< iMES

Figs. 114, 115. Pole views of monasters, first maturation mitosis.

Fig. 110. Lateral view of the same stage, showing all the chrorualin elements.

Fig. 117. Pole view of spermatid at the close of the second maturation division.

Protenor belfragei, Figs. 118-141.

Fig. 118. Spermatogonic nucleus, early prophase.

Figg. 119-123. Spermatogonic monasters, pole views.

Fig. 124. Nucleus of first Spermatocyte, synapsis.

Figs. 125-128. Lateral views of the large chromosome x, from nuclei in the late synapsis stage.

Figs. 129, 130. Nuclei of first sperrnatocytes, telaphase.

Figs. 131-134. Nuclei of first spcnnatocytes in successive prophases.

Figs. 135, 136. Lateral views of successive monaster stages, first maturation mitosis, all the chromatin elements

shown in Fig. 135.

Fig. 187. Pole view of the stage of Fig. 136.
Fig. 138. Lateral view of the anaphase, first maturation mitosis.

Fig. 139. Pole view of second spermatocy te, chromosomes not definitely arranged in the equator of the spindle.
Fig. 140. Lateral view of auaphase, second maturation mitosis.
Fig. 141. Still later anaphase.

Plate IV.

Cymus auguslatut, Figs. 142-144.

Fig. 142. Lateral view of monaster, first maturation mitosis.

Fig. 143. Pole view of one daughter cell (second spermatocyte) of the dyaster stage of the first maturation mitosis,

the univalent chromatin elements seen laterally.
Fig. 144. Pole view of one chromosome plate, metakinesis of first maturation mitosis, chromatin elements seen on

end view, except the one marked a.

Ichnodemus falicug, figs. 145-148.

Fig. 145. Spermatogonic mouaster, pole view.

Fig. 146. Nucleus of first spermatocyte, late prophase, showing all the chromatin elements.

Figs 147, 148. Pole views of monasters, first maturation mitosis, in Fig. 148 two of the chromosomes virwnl
laterally.

PcUoptlta abbreaiata, Figs. 140-151.

Fig. 149. Spermatogonic monaster, pole view.

Fig. 150. Pole view of monaster, first maturation mitOMv

Fig. 151. Lateral view of same stage, two of the large chromosomes not shown.

tKiiiiii,;il.n dortalit, Fin*. l5,'-i~>ft.

'Fig. 152. Spermatogonic nucleus, rest stage.
Figs. 153, 154. Spernmtogouic monasters, pole views.
Fig. 155. Nucleus of first spermatocyte, rest stage.
Fig. 156. Pole view of monaster, first maturation mitosis.

Fig. 157. Lateral view of the same stage, three of the chromosomes not shown.
Fig. 158. Lateral view of the anaphase, first maturation mitosis

Oneofrltut fancialu*, Fig>. 159-171.

Fig. 15!t. Spi.Tinatogouic nucleus, early prophiisc.
Fig. 160. Spermatogonic monaster, pole view.

OF THE GKRM CELLS OF METAZOA. 23")

Figs. 161-165. Kuclci of first spermatocytes, rest stage.

Fig. 166. Nucleus of first sperniatocyte, late propkase, showing all the chromatiu elements.
Figs. 167, 168. Pole views of mouasters, first maturation mitosis.
Fig. 169. Lateral view of same stage, five of the chromosomes not shown.
Fig. 170. Lateral view of anaphasu, first maturation mitosis.

Fig. 171. Pole view of second spermatocyte, chromatin elements not definitely arranged in the equator of the
spindle.

Leptopterna dolabrata, Figs. 172-116.

Figs. 172-175. Nuclei of first spermatocytes, growth period.
Fig. 176. Polo view of monaster, second (?) maturation mitosis.

Calocoris rapidus, Figs. 177-188.

Fig. 177. Spermatogonic monaster, pole view.

Figs. 178-180. Nuclei of first spermatocytes, telaphase.

Fig. 181. Oblique lateral view of the spindle before the chromosomes are arranged in the plane of the equator, first

maturation mitosis.

Figs. 183-184. Oblique lateral views of monasters, first maturation mitosis.
Figs. 185, 186. Pole views of the same stage.
Figs. 187, 188. Pole views of monasters, second maturation mitosis.

Pcecilocapsus lineatus, Figs. ISO, 190.

Fig. 189. Nucleus of first spermatocyte, early telaphase.
Fig. 190. Pole view of monaster, first maturation mitosis.

Plate V.

Pmeiloeapsus gonipJiorus, Figs. 191-10S.

Figs. 191-195. Nuclei of first spermatocytes, rest stage.

Figs. 196, 197. Pole views of mouasters, first-maturation mitosis.

Fig. 198. Lateral view of the same stage.

Phymata sp., Figs. 100-203.

Fig. 199. Spermatogonic nucleus, rest stage.

Fig. 200. Spermatogonic monaster, pole view.

Fig. 301. Nucleus of first spermatocyte, telaphase.

Fig. 202. Pole view of monaster, first maturation mitosis.

Fig. 203. Lateral view of same stage.

Coriscus ferns, Figs. 204-206.

Figs. 204, 205. Nuclei of first spermatocytes, telaphase and rest stages respectively.

Fig. 206. Pole view of first spermatocyte, the chromosomes not definitely arranged in the plane of the equator of I lie
spindle.

Acholla midtispinosa, Figs. 207-211.

Fig. 207. Spermatogonic monaster, pole view.
Figs. 208, 209. Nuclei of first spermatocytes, early prophasc.
Fig. 210. Pole view of monaster, first maturation mitosis.
Fig. 211. Pole view of monaster, second maturation mitosis.

236 MONTGOMERY A STUDY OF THE CHROMOSOMES OF THE GERM CELLS OF METAZOA.

Sinea diadema, Fiijs. SIS-SIS.

Fiv>. '-12, 213. Nuclei of first spermatocytes, tclaphase.

Fji;s. '.214, 215. Pole views of uionasters, first maturation mitosis.

Fig. 216. Lateral view of the plurivalent chromosome of the first maturation mitosis showing the mantle fibre

attachments.
Figs. 217, 218. Lateral views of successive anaphases, first maturation mitosis.

Limnobates lineata.
Fig. 219. Nucleus of first spermatocyte, rest stage.

PrioniAus crixintm, Figs. SSO-SS5.

Figs. 220-222. Spermatogonic nuclei, rest stage.
Figs. 223, 224. Spermatogonic monasters, pole views.
Fig. 225. Nucleus of first spermatocyte, rest stage.

Milyas einctui, Figi. 226-S2S.
Figs. 226-228. Nuclei of first spermatocytes, rest stage.

Hygotrechus gp., Figs. SS9-SS1.

Fig. 229. Spermatogonic monaster, pole view.

Fig. 230. Nucleus of first spermatocyte, synapsis stage.

Fig. 231. Pole view of monaster, first maturation mitosis.

LimnotrecTiut marginatu*, Figs. SSS, SSS.

Fig. 282. Nucleus of first spermatocyte, rest stage.

Fig. 233. Pole view of monaster, first maturation mitosis.

Pelocoris femorata.
Fig. 234. Spermatogonic monaster, pole view.

, Zaitha sp., Figs. XSS-SSS.

Fig. 235. Spermatogonic nucleus, rest stage.

Figs. 236, 237. Spermatogonic monasters, pole views.

Fig. 238. Pole view of monaster, first maturation mitosis.

Transactions Am. Philos. Soc. N. S. XX.

Plate IV.

,<IU

-4.z

M

HMt'^Y^"N.

.^-,*f

-KM

^-r^-vz M

f ' i

,-*/

22,

l

17.

.C,Af#.

Httt.

M2

t

MM;'

..N.

'*

It

*

20.

*

...-I/.Z

x.vi-

H.Z-"?,
CM- 3K

9 .

:';

HI.

..N.Z

36.

.

M

CM \.

*** . --l"' 1 ' Jfo M{,

i mf - 4

..N.Z

*^

d.Mt.
33. n. '*

^

-^w

--/v,
^.

^

#

'HI

Hr\

,H.l
/ ,-CM

"

,/v.

f./lftf

^/

rrr^

VMsg.

30,

,2
*

/v.

:? x7 / g'.-/V.M/

tl

..CM

'

?. -k-2--'

A(^' ' ^^M CM-'

HI

<$*

S&r

v/ ^.^

c'.WA ...Mi
* v

Montgomery, Del.

Plate i.

-

Transactions Am. Philos. Soc. N. 3. XX.

Plate V.

..C.M-f.

*

**

A,-

.*

d

-CM

a

71,

CO.

.-KM
70.

M...

wi-

Tt.

.CM

76.

?.

55:

/K

.;

*

, n,f

ft.

% -
*

fX

,

90.

\

'-H.l

N.

.(.Mi

n.

N.

fKL
f&

t.Hl.

**

Montgomery, Del.

Plate a.

Transactions Am. Philos. Soc. N. S. XX.

Plate VI.

CM CM-.. etih '

* v ^ , . : '

V> x * f *

#/ . '

Y6 - <W /

n

N.2 tt

* ./K/J// "4 ^

..,*/ </, '"ft|t|f Ir *

% *
*

#, dW /^/. /w - ^-^/

.,* ,v.w/

*> '%

* *

r yStyu ,i*. Mf ' ^

*i /^
. yr*,-^

o/i! >#
!\ v r -/KZ

-***^' ^^ "~-t(i M. f A/I "

. ^ . - "^. A Vr. /v* //7/T "'^

A/ '7 "' * ^^

" ' ^ \ A/ 'I J

///. ** /
fiKl--

//^. f-

JV> ,d.Ml f f

^ ^ , ^:

,/ o

^, ' ,^ ^^2 si* ^ ;; ^ :: :

* Ji** K ' *2I. * ; ^'

v# %.-* ij e. *

/?2 * *2- ^^ "^/^^ y. -
'> g-^'

/>:- ?*

y, K I2lf.. K* .M

//7/" - */ 1
' ^' ^ * fiiX

/ /? *?

^>

(3,> K ^5

^. ' r '^ ' /'*
-MZ

^^ X , M7
Ji .-' L.R^

a 4w. /.

S '- O-.

x, x "^^

/ ^.>, \ .
'"" ///:' ' a?-

.AjV^_5 ; l ^/^.

X-'HwV^-^ M<-~j\ %

,... ^ '" u '

9*ltf*. z -f^V-K.Z * %*

^ "' 4 r

't^ *

/K2 "" ** /^ ^? j'M A
/?^, "'^^ ^' /3j.,

a ^~~^ "*.... -" CM X -^^

"% -f./f/ ^. ,--**- , \K. M

, V- rrv -E** v

/^

/ 37, /3$r. MZ

;.

-*" X '-,v/- *'V ^

*'* C
/^*

Montgomery, Del.

Plate 3.

Transactions Am. Philos. Soc. N. 8. XX.

Plate VII.

,/u

I If. 3.

a

/ m

N,l

f:

li'T

Iff.

iff

.A/.

H. "

/a.

r

4

i a

Hi?

cl

*

m.

m.
-CM

*'

'

t.
;*

///,

196.

m-

171.

-CM.

-t.

ML-

V.2.

i,i

Montgomery, Del.

Plate 4.

Transactions Am. Philos. Soc. N. S. XX.

Plate VIII.

4* IKi

" ;-" /

^,.

in

200.

-CM

Idfis

201.

2/6.

N.Z

iff,

N.

N.2

ML

22(>.

fTT

', '; i

227.

@

> /v.

tor.

If.

233.

* M

''

2/Z.

-r -f, -

*

-2/7

*

*>

t- *-
"

"M

' n

if:'

^/<7.

.HJHU.

CM-

N..

231.

v?

-

237.

JOti

l

.Hi

2/9,

2ZI.

Montgomery, Del.

Plate 5.