Na gedurende eenige jaren wetenschappelijk werk in de tropen
verricht te hebben, is mij het voorrecht te beurt gevallen opnieuw
in de spheer van het Universiteitsleven opgenomen te worden.

In het bijzonder verheugde het mij dat ik bij mijn terugkomst U,
Hooggeleerde Went, nog mocht aantreffen in het instituut waar ik,
onder Uw leiding werkende, het mooiste deel van mijn studietijd heb
doorgebracht. Bij het voltooien van dit proefschrift dank ik U voor
al hetgene dat ik van U en door U heb geleerd.

Hooggeleerde Pulle, hoewel U indertijd in mij geen dankbare leer-
ling hebt gehad, kunt U ervan overtuigd zijn, dat dit thans wel het
geval is. Ik beschouw het als een gelukkige omstandigheid, dat een
gedeelte van dit proefschrift op plantensystematisch gebied ligt.

Hooggeleerde Jordan, nooit zal ik de voor mij belangwekkende
tijd vergeten, die ik met experimenteeren in Uw laboratorium heb
doorgebracht, zooals ook Uw colleges bij mij in dankbare herinnering
blijven.

Hooggeleerde Nierstrasz, Uw colleges zal ik mij steeds herinneren
als de kleurigste uit mijn studietijd.

Hooggeleerde Koningsberger, U dank ik voor datgene wat ik van
U tijdens Uw assistentschap in mijn eerste studiejaren mocht leeren.
De gastvrijheid welke ik thans in Uw laboratorium ondervind, stel
ik op hoogen prijs.

Hooggeleerde Honing, Hooggeachte Promotor, dat genetisch en
cytologisch onderzoek zoo aantrekkelijk voor mij geworden is, vindt
vooral zijn oorzaak in het voorbeeld dat U Uw leerlingen steeds
geeft, in het bijzonder waar het werklust, nauwkeurigheid en zelf-
critiek betreft. De tijd dat ik als Uw assistent werkzaam was is
daardoor voor mij uiterst leerzaam geworden. Ik dank U voor al
het goede dat ik van U heb ondervonden.

Hooggeleerde de Vries, voor Uw steun en belangstelling tijdens
mijn eerste jaren aan het Proefstation voor Rubber te Buitenzorg,
ben ik U zeer erkentelijk.

Waarde Bremer, dat ge mij tijdens het onderzoek voor dit proef-
schrift zoo onbeperkt hebt laten profiteeren van Uw groote kennis
en praktische ervaring op cytologisch gebied, stel ik op zeer hoogen
prijs.

I wish to tender my most sincere thanks to Mr. J. K. Spearmg for
his careful reading and correcting the manuscript.

Voorts betuig ik mijn dank aan allen, die mij bij het samenstellen
van deze publicatie op eenigerlei wijze behulpzaam zijn geweest.

From a cytological point of view Hevea may be called a practically
new subject. So far only two investigators have occupied themselves
to a certain extent with the cytology of Hevea. The first of these was
Heusser (1919) who, while carrying on investigations on the repro-
ductive organs of H. brasiliensis devoted some attention to meiosis.

As Heusser intended to approach the improvement of the rubber-
tree by means of artificial crossing, he rightly thought a preliminary
study of the reproductive organs to be of importance.

In his description of meiosis, however, Heusser does not enter into
particulars; starting from a few simple drawings he mentions the
most striking stages with a brief discussion of their connection with
the phenomena of heredity. For this investigation young buds from
random seedlings of ordinary estate-rubber were used.

Some years later there appeared a note on the number of chromo-
somes of some species of Hevea by Bangham (1931), in which the
number n = 8 reported for H. brasiliensis by Heusser was corrected.
For all the species investigated Bangham finds 2n = 34 and, more-
over, n = 17 for iJ. brasiliensis. His pubhcation contains no further
information as to the process of meiotic division.

In the meantime artificial cross-pollination has been generally
apphed to rubber-selection in the Netherlands East Indies. I have
also been engaged in this work for some years, during which I made
a number of observations which induced me to an investigation
into the germinative power of Hevea pollen (Ramaer 1932).

In the course of this investigation I found, besides a number of
normally fertile forms, a few types with degenerated pollen, which
invited cytological investigation.

After a preliminary test of normal H. brasiliensis I came to doubt
the correctness of the number of chromosomes n = 17 as stated by
Bangham.

A comparison with other tropical cuhivated plants, such asSaccha-
rum, Nicotiana, Gossypium, etc., emphasized the paucity of our
knowledge of the cytology of Hevea.

It therefore appeared to me, that a thorough cytological investi-
gation was fully justified.

The results of this investigation, which was for the greater part
carried on in the Botanical Laboratory at Utrecht, are recorded in
the following pages.

For the investigation of meiosis in Hevea brasiliensis Muell. Arg.
three clones, which were in every respect normal, were selected from
the observation collection of the „West Javaquot; Experimental Station
at Buitenzorg. Two of them, BR^ and BRj are of Sumatra origin
being obtained from the Bogor Red]oh estate in South Sumatra.
The third, Tjir. I is one of the best known clones in Java and has
originally been found at the rubber-estate Tjirandji.

Five clones were further investigated which show symptoms of
sterility: PR 104 and C.R.S. 24 from the „West Javaquot; Experimental
Station, Ct. 88 from the Culture Garden of the Department of Agri-
cuhure at Buitenzorg, KN 251 and KN 220 from the Klapanoenggal
rubber-estate near Buitenzorg.

The study of meiosis prompted a desire to include the somatic
chromosomes in my investigation. No material being at hand a
number of viable seeds of H. brasiliensis were obtained from the
Experimental Station for Agriculture in Surinam. To the director
Prof. Dr. G. Stahel I wish to express my profound gratitude for the
sending of the seeds and to Prof. Dr. L. P. de Bussy for his kindness
in acting as intermediary.

The seeds were germinated for the greater part in the hothouse of
the Utrecht botanical gardens, after which the roots of the young
seedlings were used for further investigation.

Material of Hevea Spruceana Muell. A-RG.,oiH.guianensisAvBL.
and of H. collina Huber was collected as well as of H. brasiliensis.
Of these species a number of trees exist in the Culture Garden at
Buitenzorg. I am greatly indebted to Messrs. H. de Veer and Ir. G.
G. Bolhuis for information concerning the origin of these forms and
for other valuable assistance.

H. Spruceana and H. collina were originally obtained from the
Selection Station at Santarem (Brazil) and H. guianensis from the
Selection Station in Surinam.

Finally the Culture Garden at Buitenzorg is in the possession of
some peculiar forms of Hevea, which are considered to be species-
hybrids of H. collina X H. brasiliensis (D2—78) and oiH. Spruceana
X H. brasiliensis (D2-—49). From these trees material was also
collected.

That the rubber clones cultivated at present belong to the species
H. brasiliensis Muell. Arg. is hardly ever doubted. They originated
from ordinary estate rubber in the Netherlands East Indies which
in its turn sprang from seeds gathered by Wickham in Brazil in 1876.
In connection with the numerous variations of the characters of H.
brasiliensis in the Netherlands East Indies the question of the purity
of this species has aheady been repeatedly discussed.

As is communicated by Cramer (1914) the possibility that among
the seed material of Wickham, besides H. brasiliensis, H. collina
may have occurred, is admitted by Huber, so that in the Netherlands
East Indies H. brasiliensis might have been mixed with H. collina.

According to Heusser (1919), however, the fact that among the
great number of flowers investigated, he has never met one of the
collina type, is an argument in favour of the contrary.

In my opinion it is doubtful whether as much reliance should be
put on this argument as Heusser does, in view of the fact that abso-
lutely nothing is known about cross-fertilization of the various
species of Hevea.

As I have aheady said the variation of the characters of cultivated
H. brasiliensis is very considerable. There is no characteristic that
does not vary to a very great extent. This has even been utihsed for
the drawing up of tables of identification for young trees of the best
known cultivated clones (Frey—Wyssling, Heusser and Ostex-
dorf 1932).

In Brazil similar variation is found, as is reported by Ducke (1933)
who paid much attention to the classification of Hevea, and made
observations in various parts of Brazil.

are evidently very great. In reviewing my material I think I cannot
do better than take the classification of Ducke for my guide, as the
latter is considered the greatest authority in the field of the taxo-
nomies of Hevea.

The original classification of the genus into the subgenera Euhevea
and Bisiphonia, according to the existence of either one or two (com-
plete or incomplete) whorls of stamens, is made by Mueller Argo-
viensis (1873) and also used by Huber (1905) but recently modified
by Ducke (1930). The latter distinguishes 5 groups, the fir.st of which
approximately corresponds with the subgenus Euhevea, which inclu-
des the species H. guianensis Aubl. and H. coZ^m« Huber. They are
chiefly characterised by one regular whorl of 5 stamens, the practi-
cally entire absence of a disc in the male flower, and small, hard,
leathery leaflets.

To this I may add that the forms of H. guianensis and H. collina
I investigated, when compared with H. brasiliensis, are very slowly
growing trees and produce extremely small seeds. The seeds are half
the volume of those of H. brasiliensis, and are triangular in shape.

When the trees begin to shed their leaves the latter assume brilliant
red and yellow colours; their latex is yellow and resinous.

The great similarity of these two species induced Ducke (1933)
to consider the species collina Huber a variety of guianensis Aubl.
For conveniences sake, however, I shall go on speaking ol H. colhna.

The 2nd, 3rd and 4th groups include five species which possess
an irregular number of stamens in two indistinctly separated and
incomplete whorls (See below Ch. 3). They belong to the subgenus
Bisiphonia of Mueller Argoviensis.

In the 5th and largest group Ducke includes H. brasiliensis Muell.
Arg. and H. Spruceana Muell. Arc. Both species are characterised by
two complete and regular whorls of 5 stamens and therefore also
belong to the Bisphonia type of Mueller Argoviensis. Besides in
the number of stamens they differ from the first group by the pos-
session of a disc in the male flower, large leaves which are scarcely, or
not at all, leathery, more vigorous growth, larger seeds, and white
latex. The leaves, at the moment of falling, are pale yellow or pink.

H. brasiliensis always has yellow flowers, H. Spruceana is, as far as I
have been able to ascertain, the only Hevea with purple flowers.
Each of the 3 outsides of the pericarp is concave with H. brasiliensis,
convex with H. Sprttceana. The fruit of H. brasiliensis bursts open
spontaneously, the seeds and parts of the pericarp being hurled
away; the fruit of H. Spruceana opens slowly, the seeds dropping
vertically and the parts of the pericarp remain hanging. The fruit,
moreover, is oblong and the seeds are approximately twice as long
as those of E. brasiliensis, whose fruit is just as long as it is wide.

in the two species (fig. 1); the flowers of H. Spruceana are long-
haired, those of H. brasiliensis are short-haired. The buds of H.
brasiliensis taper, those of H. Spruceana are blunt, and finally the
leaves of the former are glabrous, those of the latter hairy.

The differences between the two species are much more consider-
able than the variations within the species brasiliensis.

As regards the specimens I investigated they entirely agree with
Ducke's description, so that I can be quite sure about their identity.

Of the five sterile forms of H. brasiliensis it can also be stated
that they show all the characteristics of H. brasiliensis with the
always present variation.

Fig. 1 shows the various shapes of the disc-lobes in male buds
of all forms mentioned; they are found either against the foot of the
column (shown to the left) or more on the receptacle.

As to the H. brasiliensis seeds obtained from Surinam, they had
all the characteristics of brasiliensis seeds, and their seedlings resem-
bled also those of H. brasiliensis, but more cannot be said.

In § 1 a few Heveas were already mentioned which might be looked
upon as species-hybrids. Up till now however, these trees have not
been investigated, so that a short description follows below.

D2—78 is looked upon as a hybrid of H. collina X H. brasiliensis.
Its growth is satisfactory and it has leaves which are fairly large and
not leathery, therefore corresponding rather with those of H. brasi-
liensis. In the male flowers a disc is found (fig. In) whose 5 lobes are
smaller than those of H. brasiliensis. The flowers are of an intermediate
size, form and colour, the seeds resemble more those of H. brasiliensis.

The number of stamens varies from 5 to 10. Moreover they are
mostly very irregularly inserted in two whorls which are not marked-
ly separated. Half- or badly developed stamens also occur.

Of this three I have investigated 151 flowers; table I shows the
numbers of stamens occurring in the flowers examined. Half
stamens have been counted as complete ones.

Taken as a whole, therefore, the number of stamens is intermediate
though tending in the direction of the greater number. The same
can be said about the other characteristics, which are nearly all
intermediate with a slight approach to the brasiliensis type.

sinensis. The coloration of the flowers is striking, these showing a
purple basis with yellow lobes.

The shape of the fruit is intermediate with a slight approach to
the brasiliensis type. I have not been able to observe their bursting;
from the state of some dry fruits received from Buitenzorg, however,
the conclusion that this characteristic is also intermediate is j ustified;
it is supposed, however, that the seeds are not hurled any consider-
able distance as the valves of the fruit remain partly together and
attached to the stalk. The seeds are very big, being almost the size
of Spruceana seeds, but somewhat thicker, not broader.

Markings and colour of the seeds are intermediate. The shape of
the disc of the male buds is intermediate but varying in several
flowers (fig. Ik, 1), the hair of the flowers and the shape of the buds
are intermediate, the leaves are practically glabrous.

Further it can be stated that the Herbarium at Utrecht possesses
herbarium material of a tree collected by Ducke in Amazonas,
which is thought by him to be a hybrid between H. Spruceana X
H. brasiliensis. The flowers of this plant are also predominantly purple
with yellow lobes, while the leaves are glabrous. The description
of these „specie.s-hybridsquot; brings up the question of hybridization.

Experience in the field of the hybridization of Hevea is limited
to the species brasiliensis. The results of artificial cross- and self-
fertilization enable us to form an idea of the fertility of cultivated
H. brasiliensis. Literature on this subject can be found with s' Jacob
(1931).

It appears then that, just as all other characters, fertility varies
a great deal.

As a rule cross-fertilization is considerably more successful than
self-fertilization. During the years 1928—1931 I myself carried out
40609 cross-fertilizations with an average success of 6.7% against
3250 self-fertilizations of which 0.9% were successful. In 1931 I
began observations and experiments on interspecific hybridization,
induced by the presence of the „species hybridsquot; in the Culture
Garden of Buitenzorg, as described in § 2. These plants were not
obtained by artificial cross-pollination but according to my infor-
mant, Mr. H. de Veer, they originated as follows.

A number of seeds were collected from two trees, of H. collina and
H. Spruceana respectively, at the Bogor Redjoh estate and were
germinated in two separate beds. The seedlings of each of the beds
soon separated into two entirely different groups, one showing a
striking resemblance with the mother tree, the other group showing
brasiliensis influence.

It was reasoned that the seeds of the first group resuUed from
self-fertihzation and those of the second from cross-fertilization by
pollen from the surrounding H. brasiliensis; so the latter group
would, in this case, be species-hybrids. A number of these plants
planted out in the Culture Garden at Buitenzorg, grew up well. In
1931 they were 6 years old, flowered and produced fruits and viable
seeds.

In the same year I myself repeated this experiment for check. Of
each of the trees described in § 2: H. guianensis, H. collina and H.
Spruceana seeds were collected and germinated in three separate
beds. The progeny of H. guianensis for the most part died but that
of the two other forms grew up well. After a few months striking
differences were visible between the plants in both of the beds.

Two groups were clearly distinguishable. In the coUina-heA part
of the plants dropped considerably behind the other part in growth.
After six months the latter were as much as from 2 to 3 times as tall
as the former. The smaher plants showed, in their leaves, typical
characteristics of H. collina; the taller more of H. brasiliensis.

With the Spruceana-gTou-p the same was observable, although the
growth differences were less striking than in the progeny of H. collina.
Although the differences in the appearance of the foliage and other
habit characters of immature plants have not been discussed in
the preceding sections, they are great enough to enable the above-

Finally I reciprocally crossed H. brasiliensis with H. Spruceana.
H. brasiliensis x Spruceana proved very successful, better even
than many crosses between clones of H. brasiliensis.

Fertilization took place, but the young fruits were all shed, a
phenomenon which also frequently occurs with H. brasiliensis and
is attributed to the physiological condition of the plant.

zation of H. guianensis and H. collina as flowers of neither species
were available during the flowering period of H. brasiliensis.

During an investigation of the germinative power of pollen of H.
brasiliensis (Ramaer 1932) I found with clone P.R. 104 a great deal
of dead pollen besides a small quantity of full pollen grains, whose size
varied considerably; only a few of them germinated on sugar-agar.

The male flowers and stamens were externally normal but at
closer investigation most anthers were filled with dead pollen. One
single loculus had almost exclusively full pollen grains. Degeneration
of pollen appeared to take place after the tetrads had fallen apart.

R.P. 104 produces exceedingly little fruit so that this clone can
almost be called sterile.

During experiments concerning artificial pollination on the
Klapanoenggal estate two clones KN 220 and KN 251 were found,
the male flowers of which did not attain to full development but
were shed in the budding stage. When the buds were opened there
proved to be, on the place of the column with the 10 anthers, a hard
little pin, which often lay loose on the receptacle.

A closer examination of the male buds revealed that here not only
degeneration of the young pollen takes place but also of the whole
stamens. It was remarkable that a few male flowers could be induced
to open by removing all other buds from the panicle.

As to the fruiting of these clones, KN 220 never produced any
fruit; KN 251, on the contrary, produced a great many fruits and
viable seeds.

On going through the observation collection of the „West Javaquot;
Experimental Station I discovered 2 more clones, which behaved
in every respect like KN 251.

These three clones therefore may be looked upon as female plants
as pollen is never produced. As I already remarked in my publication
above mentioned, such forms are eminently fit for experiments on
natural insect-and wind-pollination.

Finally, during the cytological investigation described in Ch. 2,
it moreover appeared that the „species hybridquot; H. Spruceana X
H. brasiliensis also produces dead pollen and is consequently male
sterile.

Meiosis was almost exclusivelj^ investigated in pollen mother-cells;
the technical details mentioned in this section consequently refer
to male bud preparations. Some additional observations, however,
were made on somatic chromosomes. Root tips of young seedlings
were fixed in Bouin-Allen B—15 or in La Cour's 2BE (La Cour 1931)
and stained with Newton's gentian violet.

The young male buds were collected in the neighbourhood of Bui-
tenzorg, in the flowering season of 1932. As a rule the moment of
fixation ranged from 9 to 11.30 a.m. For fixing Carnoys fluid (acetic-
alcohol) was always used. The fixed material was carried to Europe
in paraffin and handled in the Botanical I^aboratory at Utrecht.

Being fixed with Carnoy the one stain was Heidenhain's haema-
toxylin. A preliminary test had already proved this stain to be satis-
factory. Yet the results obtained later on appeared very unequal,
and on the whole only moderately satisfactory. E.g. Hevea Spruceana
caused great difficulties with this stain and I have not really succeeded
in obtaining a completely succesful preparation, the plasma almost
invariably retaining too dense a colour. Other types proved to be
very variable in this respect. Staining with haematoxylin with sub-
sequent differentiation with a saturated solution of picric acid in
water (Hsu Chuan Tuan 1930) brought no improvement.

Gentian violet by Newton's method stained the chromosomes very
unsatisfactorily. Previous treatment with chromic acid (Clausen
1929, Skovsted 1934), occasionally yielded better resuhs, but Hdden-
hain's haematoxylin remained the superior method so that eventu-
ally it was applied throughout the investigation.

Although the cytoplasma generally remamed deeply stained and
the chromosomes sometimes too lightly, the nucleolus was often very
faintly stained and not infrequently entirely transparent. It was
remarkable that with some types (e.g. Hevea guianensis) the nucleolus
always remained pitch black.

At first sections of 15 ju, thickness were made, but these were after-
wards proved to be too thin for prophase nuclei. 18 [i. is the most
suitable thickness for Hevea.

Longitudinal sections are to be preferred to cross ones as the
contents of the anthers are better surveyable, first in differentiating
and afterwards in studying the meiotic stages. In the cross section
of a loculus two or three pollen mother-cells at most are found
whereas in longitudinal sections often much more, sometimes as
many as from 12 to 16.

A macroscopic expedient for the selection of buds in which the
meiotic divisions are in progress, as e.g. the „benderaquot; stage of Sac-
charum (Bremer 1921) and the macroscopically measurable length
of the buds of Oenothera (Geerts 1909), is not known for Hevea. Nor
do Heusser (1919) and Bangham (1931) supply any indications.

The fixed material therefore consisted of buds of much varying
lengths; for which reason a great number of longitudinal sections
were made first, for purposes of orientation. With the aid of ocular-
and stage-micrometers the lengths were determined of those buds in
which stages of meiosis occurred.

These lengths appeared to show considerable differences for the
various forms. The figures of Table II will serve to illustrate
this.

BR 2 2.04—2.21
Tjir. 1 1.79—1.96
PR 104 1.70—1.87
KN 251 1.70—1.87
Ct 88 1.70—1.87
C.R.S. 24 1.70—1.87
KN 220 1.70—1.87

H. collina X brasiliensis. . . 1.11 —1.28
H. Spruceana X brasiliensis . 1.28—1.45
At the time that meiotic division is taking place the female buds
are much bigger than the male, one female bud of Ct 88 in which
pachytene, diakinesis and telophase were found, having a length
of more than 3 mm.

The drawings were made in the usual way with the aid of a drawing
apparatus. Those chromosomes lying in low focus are indicated by
a lighter shade in the drawings when distinction is necessary.

Fig. 2. Somatic divisions from root tips of Hevea brasiliensis.
a. Fixation with La Cour's 2 BE, b. with Bouin-Allen B-15. 4000 X .

chromosomes. If the magnification of about 4000 X is taken into
consideration, they appear to be very small. The drawings were
first made at a magnification of 2700 X and afterwards enlarged
11/2 times in copying.

Plate 2a is from a root tip fixed with La Cour's 2 BE, plate 2b
from a root tip of another seedling and fixed with Bouin-Allen
B—15. The difference in size between the chromosomes is consider-
able, those of b are more than 1V2 times as long as those of a. Whether
the latter is due to different fixation or to different nutritional con-
ditions of the root tips, cannot be said.

As regards the structure of the chromosomes, 2a is more satisfactory
than 2b. A constriction is clearly visible in various chromosomes,
in b, however, not at all. This is a result of the different influence
of the two liquids used for fixation. It is emphasized by Huskins and
Smith (1935) that the structure of chromosomes appears unsatis-
factory, when fixing liquids are used which contain ureum.

Slides from growing points have also been made and stained with
Newton's iodine gentian violet ; good division figures however were
not found.

Frequently, somatic chromosomes in the tapetal cells of the anthers
could be observed in which a constriction was visible, even with
Heidenhain's staining. The tapetal chromosomes are evidently bigger
than those in the root tips. Counts, however, were not possible, no
regular metaphase plates occurring.

The pollen mother-cells in various anthers of a Hevea flower are
often in different stages of meiotic division. In differentiating this
is a disadvantage, but a convenience in the study of transitions.
According as a stage is of longer duration it is found in a greater
number of stamens at the same time.

Neither is the same stage always found, in the 4 loculi of the same
anther and it even occurs that in the upper part of a compartment
pollen mother-cells have reached a more advanced stage than those
in the lower part. This happens at the end of diakinesis and also
occasionally at the beginning of diplotene.

After the resting stage following the last somatic division, the
prophase of meiosis begins with a leptotene stage, which is indistinct
and very complicated.

The pollen mother-cells are still condensed, and fill the locuh
almost completely. (For the structure of the flowers and stamens see
Heusser 1919). Either simultaneously with, or possibly after the
beginning of zygotene, which is just as indistinct as the leptotene,
synizesis sets in. Cells with a synizetic knot are already detaching
themselves from each other and from the tapetum, while the angles
of the cell-walls are rounding off.

This continues during the subsequent pachytene stage, the first
distinct stage of prophase.

Pachytene now passes into diplotene, which hke leptotene consists
of a comphcated mass of very thin threads and loops. The cells have
now become entirely rounded off and lie just touching, in a row in
the middle of the loculi.

In the diplotene stage the contraction of the chromosomes sets in,
which process continues until metaphase. Diplotene gradually passes
into diakinesis, when the contents of the nucleus spread over the
nuclear-membrane. After the latter and the nucleolus have dis-
appeared a contraction of the chromosome group occurs. Next the
chromosomes spread among the spindle fibres (which in the mean-
time have made their appearance) and then range themselves in the
metaphase plate.

Metaphase and anaphase show a fairly normal course and pass
off rapidly. In the telophase groups alveolisation takes place and
soon two additional nuclear-membranes are formed.

This is followed by interphase (interkinesis) in which each separate
chromosome becomes clearly visible, while one or several nucleoli
appear.

Between the two nuclei no cell-wall is formed. After the disap-
pearance of the two nuclear-membranes metaphase, anaphase and
telophase of the second meiotic division take place.

After the telophase-II follows a stage which greatly resembles the
interphase. The now very small chromosomes are again separately
visible until the pollen mother-cells break up into tetrads of young
uni-nucleate pollengrains, which soon round off.

Leptotene, zygotene and synizesis. Very gradually the granular
contents of the resting nucleus (fig. 3) change into an intricate net-
work of very fine threads in which the chromomeres are sometimes
clearly visible as small granules; the threads, however, cannot be
traced (fig. 4).

Anastomoses soon appear, but it is not possible to find any indi-
cations of pairing having begun. The chromatin is in the form of thin
and thick threads intertwined. That the thicker parts have a double
nature is probable but has not been proved. It is impossible to make
out a distinct zygotene stage in Hevea.

Moreover at this stage the contents of the nucleus usually show
contraction; the chromatin in the form of a dense tangle, together
with the nucleolus, hes against the nuclear-membrane. This synizesis
was repeatedly met with in the preparations, as in most plants
where the fixation has not been done with the most painstaking care.

Most cytologists [see Belar (1928), Darlington (1932), Sharp
(1934)] at present assume synizesis to be an artefact which may
occur at late leptotene as well as at zygotene or early pachytene,
the aspect of the knot varying in each of these cases.

In Hevea synizesis usually occurs at zygotene. Fig. 5 illustrates a
case in which two threads that are just pairing project from the main
mass; thus it is evident that in Hevea parasynapsis takes place. Four
pairs of chromomeres are clearly visible. The nucleolus in this nucleus
is practically colourless.

Sometimes the early pachytene shows contraction when thick
ends of thread are projected from the knot.

Often the degree of contraction varies in the nuclei of a single
loculus, some nuclei showing no synizetic contraction at all, others
to a certain amount, others again showing complete contraction.

On the whole it seems improbable to me that the dense synizetic
knot should offer a favourable opportunity for the side-by-side
pairing of the chromosomes which are very long at this stage. The
assumption that synizesis is an artefact has every probability on
its side.

ding I have not yet succeeded in analyzing any pachytene nucleus.
The nucleus is filled with a confused mass of thick chromatin threads
which cross and recross in every direction (Fig. 22a). There is no
question, however, of a continuous spireme for, although it is only
rarely possible to trace any one chromosome completely, a large

number of free ends may be observed. In one case I counted as many
as 18; the chromosomes are doubtless present in the haploid number,
but mostly stick together. Often a chromosome seems to be attached
with one end to another chromosome in an arbitrary place; it also
happens at times that the thread appears broken and, especially at
late pachytene, that very often ends of threads lie exactly opposite
each other. These peculiarities, however, cannot possibly be solved
with the available material.

Diplotene. The double nature of pachytene threads does not be-
come recognizable again until the transition into diplotene. The two
partners begin to separate in several places; at first thickenings
show in these points, but soon after openings become visible. These
by extending approach each other, until the double chromosome
consists of a succession of loops, connected by means of the chiasmata.

This process does not start simultaneously in all chromosomes nor
does the separation of the double thread in all points begin at exactly
the same moment. Sometimes the first opening becomes visible in
the middle of the unseparated thread, but frequently at the end.

This is about all I could
observe with the staining
applied. Only a very few
Inbsp;chromosomes were to be

Although diplotene is
just as complicated as
leptotene, the appearance
of the threads is entirely different. Chiasmata are present in large
numbers (fig. 7, p. 209) but the maximum number per chromo-
some I have not been able to ascertain. The diplotene is also very
sensitive to Carnoy-fixation, some contraction of the contents of the

nucleus is frequently found so that the thin threads stick together
in a great many places, and it is impossible to trace them.

At diplotene contraction of the chromosomes sets in which is
accompanied by a shifting of the chiasmata. It is not until contraction
has shortened and thickened the chromosomes, that they become
at all clearly visible in their entirety. Yet I have been able to see
even at late diplotene only a few separate chromosomes clearly.

After the transition to diakinesis there is a definite improvement.
Fig. 8 shows the cut portion of a nucleus in which the chromosomes,
according to my estimate, have been shortened to about half their
original length. It is not yet possible to count them at this stage.

Diakinesis. Fig. 9 represents diakinesis in H. guianensis. Here
18 pairs of chromosomes can be counted with certainty. The nucleolus
which is actually pitch black, had to be left white for the sake of
clearness. The chromosomes are already in an advanced stage of
contraction, but the greatest contraction is reached after diakinesis.

The nuclear-membrane disappears just before the nucleolus. On a
few occasions I have noticed instead of one large nucleolus several
small ones which are never observed at diakinesis proper.

Moreover a certain contraction of the contents of the nucleus takes
place in consequence of which the 18 bivalents cluster together
(fig. 10).

Terminalisation of chiasmata. Although in this first investi-
gation I have not been able to tell the chromosomes of Hevea apart and
to describe them separately, yet I have made a number of observa-
tions which may be recorded on their form and their transformation
from pachytene to diakinesis.

As was already shown in fig. 6 a number of loops and chiasmata
make their appearance at early diplotene, occuring at fairly regular
distances in the double thread. The number in this case including

the terminal chiasmata is four, but in late diplotene the chiasmata
no longer occur at regular intervals.

One (sometimes two) of the openings widens at the expense of
the others. Owing to the widening of these openings the chiasmata
are shifted along the chromosome and become crowded together,
whilst they gradually move towards one or both ends of the chromo-
some. This process can take place in various manners as is shown
in fig. 11. In some cases a terminal opening widens, in others a more
central one. In the former case the chiasmata are all shifted in the
same direction; in the latter in both directions.

This difference is possibly connected with the first appearance of
the openings at early diplotene. The opening which become visible
first, might continue till diakinesis.

The shifting chiasmata, after becoming aggregated, finally are
resolved at the ends of the chromosomes with the result that the
number per chromosome decreases. This process of terminalisation
(Darlington, 1929) continues until diakinesis when the bivalent,
retain a minimum of chiasmata. At diakinesis ringshaped and cross-
and V-shaped bivalents occur. The former have 2 chiasmata, the
latter one.

The number of 2-chiasmata-1) bivalents amounts to four or five,
the number of 1-ch-bivalents to fourteen or thirteen. I have been
unable to ascertain this exactly, but I got the impression that this
proportion is constant, at least in the case of H. brasiliensis. Of the
other species too little suitable material was available.

The 2-ch-bivalents may contain two terminal chiasmata, or one
terminal chiasma and one „interstitialquot;. The chiasma of the 1-ch-
bivalents may be completely terminalized but it is more frequently
interstitial, however, not median.

Only one or two 1-ch-bivalents have a median chiasma. At middle
diakinesis there are always some bivalents which have already be-
come so contracted that their shapes can no longer be ascertained.
Contraction apparently does not proceed at the same rate in every
chromosome, and possibly the same applies to terminahsation. This
makes it very difficult to decide whether the terminalisation of

a 2-ch-bivalent is already finished at diakinesis, and consequently
whether another 1-ch-bivalent might arise from it.

On the question of the splitting of the two paired chromosomes
in the prophase, so that each chiasma is constituted by the crossing
of two of the four chromatids so formed, I have made the following
observations.

At diplotene I have been unable to observe this split but the ap-
pearance of the chiasmata indicates its existence, as the upper and
lower angles of a chiasma frequently are slightly rounded off and
not distinctly acute, to be the case if the chiasmata were formed
simply by the crossing of two threads, without any others being
present.

At early diakinesis, however, a fissure is clearly visible in many
cases. Some of these cases are shown in fig. 11. Moreover the structure
of a chiasma is more or less visible here and there.

Metaphase-I. After the contraction following diakinesis, the
chromosomes scatter through the spindle, which has meanwhile

made its appearance (fig. 12, p. 211) and afterwards range themselves
in the equatorial plane.

The side view of the metaphase plate is very regular (fig. 13), but
is of short duration as pure metaphase is comparatively seldom

found, some chromosomes having usually passed into anaphase. The
spindle is frequently found close to the cell-wall.

Metaphase also shows that only a few bivalents occur with more
than one chiasma. Most of the bivalents are bar-shaped with a slight
constriction in the middle and only a few are diamond-shaped or oval.

A polar view of the first metaphase (fig. 14) often affords a favour-
able opportunity for counting. With BRj and Tjir. I {H. hrasi-
liensis) I have repeatedly, with certainty, found the haploid number
to be 18; with BRj I have never obtained a clear metaphase, on the
other hand many diakineses and interkineses allowed accurate
counting. The metaphases of H. Spruceana and H. guianensis too
were hardly distinguishable. Whilst the cytoplasma was densely
stained the chromosomes lay very close together. The number of
chromosomes of these species has also been determined in dia.kinesis.
Of H. colUna good metaphases with 18 chromosomes were found.
So the haploid number of IB occurs in all the species investigated.

Considerable variation in the size of the chromosomes was noticed
in different metaphase plates, even with flowers from the same tree.

Fig. 15. Anaphase-I, side view. 1800 x. Fig. 17. Telophase-I. IBOOx.

chromosomes may be early, others late (fig. 15). The polar view of
a few anaphases allowed both groups to be counted at different foci

Two cases of this kind are shown in fig. 16. The similarity of the
grouping of the chromosomes is clearly visible. At anaphase-I the
chromosomes are square, many show signs of the longitudinal split

(which apparently occurred in the prophase) but the halves remain
in very close contact and pass in this condition into telophase.

The chromosomes in fig. 16b show a striking difference in size
from those in fig. 16a (here the species are different and the stages not
quite identical).

Telophase (fig. 17, p. 215). Of this stage there is little to be said.
Owing to alveolisation the chromosomes themselves entirely disap-
pear, to reappear at interphase.

Interphase. For the counting of the chromosomes, interphase
(interkinesis) is at least as important as the metaphases of the two
divisions. In the first place the chromosomes may appear so clearly
separated that even with a small magnification it is possible to count
them. Not infrequently counting has been possible at interphase
when metaphases were useless or lacking.

Moreover interphase affords an opportunity to check the distri-
bution of the chromosomes at anaphase-I which is of very great im-
portance especially in the forms with irregular meiosis.

Fig. 18 shows an interphase of BR2. At this stage the chromosomes
have often a deeply cleft appearance; at the same time they show a
certain similarity of form with some bivalents at diakinesis. Besides,
the whole stage recalls diakinesis in many respects and may better
be called prophase-II than interphase.

Not infrequently more than one nucleolus is found in the inter-
phase nucleus and once I found nuclei with 5 and 6 nucleoli.

With H. guianensis, H. collina and H. Spruceana the nucleoli were
alwaj's perfectly black.

The transition to metaphase-II is also accompanied by some con-
traction of the nuclear contents just as in the first division.

A stage during which the chromosomes spread through the spindle
has not been observed.

Metaphase-II. This metaphase is distinguished from -I, by the
smaller size and the more angular shape of the chromosomes (fig. 19).
Splitting, as is observed at interphase, only remains visible here and
there by an inward bended outline of the ends.

The direction of the two spindles with respect to each other varies
considerably. They sometimes he parallel and in a few cases a polar
view allowed the two to be counted.

Again in metaphase-II many countings were done which always
yielded the same result, n = 18.

Anaph.a.se-11. Only in one single case were countings made at this
stage. Anaphase-II proceeds more regularly than anaphass-I and is
soon followed by telophase.

Telophase is followed by another stage in which the chromosomes
become separately visible, and although they are very small, coun-
ting is occasionally possible. This stage resembles interphase very

closely (except that the chromosomes show no split) and precedes
the formation of tetrads. Tetrads resuh from a constriction of the
protoplasm of the pollen mother-cell.

As stated in Ch. I, meiosis has been investigated in six forms
showing male sterility. PR 104 is almost sterile for both sexes; KN
251, Ct 88, C.R.S. 24 and H. Spruceana x brasiliensis (D2—49)
are male sterile and KN 220 is entirely sterile for both sexes.

PR 104. The appearance of the locuh in the longitudinal sections
is already abnormal at the time of the pollen mother-cehs. Whereas
in normal cases (Ch. II, § 2) poUen mother-cells are already fairly
considerably rounded off during pachytene, and separated from the
tapetum, ranking themselves in a line in the middle of the loculus,
the pollen mother-cells in this case fill up the compartment entirely
until pollen formation and are also more angular. This gives the
loculus the appearance of growing too slowly. Or, as is reported by
Rosenberg (1917) of a similar case with Hieracium: The divisions
of the pollen mother-cells may begin at a very early stage in the
development of the loculus.

Further the most divergent stages of meiosis may be found in
the pollen mother-cells of one loculus. E.g. diakineses and inter-
phases are often found lying side by side; sometimes pollen mother-
cells with 4 nuclei, second anaphases and pachytene occur in one
and the same compartment. Nor is meiosis itself normal. The
deviations which occur are also found in the following forms and
will be described together with the latter.

KN 251, Ct 88, and C.R.S. 24. In contradistinction from PR 104
the locuh with the pollen mother-cells of these clones present an
almost normal appearance until the formation of uninucleate pollen-
grains. Occasionally, however, a flower-bud is found in which the
stages of meiosis occur, but whose tissue has an unhealthy appearance
and is obviously already degenerating. The young pollen of the
„species-hybridquot; H. Spruceana x brasiliensis —49) also degener-
ates but the stamens do not.

Metaphase-I is irregular in the 4 clones and the „species-hybridquot;
mentioned above. The chromosomes do not always arrange them-

selves into an equatorial plate but frequently separate in various
places in the spindle. Fig. 20 shows a number of such irregular meta-
phases; a and b are metaphases of PR 104, c of C.R.S. 24, d and e
of KN 251, f and g of Ct 88 and h of the hybrid H. Spruceana X brasi-

Fig. 20. Metaphase-I from male-sterile forms of Hevea.
a, b. PR104. c. C.R.S. 24. d, e. KN 251. f, g. Ct 88. h. H. SpruceanaY.
brasiliensis. 2700 X .

liensis (D2—49). In these figures only the important and clearly
visible parts of the metaphase groups are drawn jet black (which
here does not indicate the chromosomes at highest focus).

some is often seen lying near the edge of the metaphase-plate (fig. 20
c and d), which is never found at metaphase in normal clones. The
assumption that it is a univalent is obvious and investigation of
diakinesis has proved this assumption to be correct; splitting of the
univalent has repeatedly been observed, (fig. 20 c). The univalent
chromosome also occurs sometimes at one of the poles while the
other chromosomes are still arranged in the equatorial plate (fig. 20a).

Besides univalents and bivalents, trivalents are also found at
metaphase, although the opacity of the mass of the chromosomes
often causes doubt. Yet I am sure that Hke univalents they occur
frequently, if not regularly. In fig. 20b (below) is a trivalent in early
anaphase, in fig. 20 e, f and h are trivalents in metaphase.

Further it is probable that per metaphase more than one univalent
and more than one trivalent occur (fig. 20 e and f). Fig. 20 g shows
what appears to be a 7-valent group, but may merely a case of the
chromosomes sticking together (fixation artefact?). Figs, g, h andi
are all drawn from PR 104-slides.

Fig. 21. Trivalents and quadrivalents from diakinesis in male-sterile clones
of Hevea brasiliensis. 2700 x .

recognize but it is often impossible to distinguish multivalents with
certainty from bivalents. Fig. 21 shov,^s a number of multivalents
at diakinesis; a, b and c are trivalents of KN 251, PR 104 and Ct 88
respectively; d is a multivalent of Ct 88 from the same nucleus as c,
but it cannot be said with certainty whether this is a trivalent or a
quadrivalent. The same applies to e and f, muUivalents of PR 104;
g and h are very probably quadrivalents and i certainly is. Figs,
g, h and i are all drawn from PR 104-slides.

In side views of metaphase I have not been able to recognize any
quadrivalents. Presumably they are of less frequent occurrence than
either trivalents or univalents.

The distribution of the cliromosomes at the anaphase cannot always
be regular. If there are 1 univalent and 1 trivalent, which may occur
rather frequently, the following cases are possible:
a. the univalent and ^/g of the trivalent go to the same pole; the

distribution then becomes normal (18—18).
h. the univalent and ^/g of the trivalent go to the same pole; the

distribution then becomes 17—19.
c. the univalent divides itself longitudinally, the halves each go to
a pole, making the distribution 18—19.

Investigation of a number of interphases mostly indicated 18—18,
twice 17—19, and only once 18—19. The number of suitable inter-
phase nuclei was too small to determine the frequency of the various
types of distribution which occurred.

The determination of the number of chromosomes in the inter-
phase nuclei is unimportant except where a clone produces good
(i.e. viable) pollen, but only PR 104 yields, besides a considerable
quantity of dead pollen, a few full grains; after artificial pollination
with this pollen, however, no fructification occurs. The anthers of
PR 104 and R. Spruceana x brasiliensis (D2—49) further remain
normal, bat do not open spontaneously. The „species-hybridquot; has
only dead pollen.

The stamens of KN 251, C.R.S. 24, and Ct 88 also degenerate. The
anthers, together with the column, become one hard shapeless pin,
which often lies loose on the receptacle.

In the megaspore mother-cells of these clones I have only found,
of ail the stages which are of importance, one early diakinesis. In
this a chromosome occurs which may be a trivalent.

KN 220. An entirely different kind of deviation fi'om normal
meiosis is found in the pollen mother-cells of this completely sterile
clone. KN 220 is an instance of an asynaptic Hevea with a meiosis
in which practically only univalents occur.

The earliest stage I have acquired is synizesis of which nothing
can be said. A typical pachytene stage does not follow, the contents
of the nucleus after synizesis consisting of thin unpaired threads.
Fig. 22 shows the principal stages of this asynaptic meiosis. Compare
the asynaptic „pachytenequot; of 22 b with a normal pachytene as re-
presented in 22 a. Though homologous chromosomes he side by side
in various places no pairing takes place.

This stage is sensitive to Carnoy fixation, a few threads always
remaining clustered together at the nucleolus. Nothing is visible but
a mass of threads partly contracted and sticking together. From
pachytene to diakinesis the nucleus has proved unsuitable for critical
observation.

Not until diakinesis does a great number of univalents become
distinguishable. One diakinesis is shown in fig. 22 c, where 33 distinct
chromosomes are drawn.

One of the black chromosomes might be taken for a bivalent and
a few chromosomes cover each other exactly. Moreover the nucleolus
has been left white in the drawing, but is actually completely black.
As a whole this stage did not prove suitable for careful counting.
Diakinesis is the only stage in which chromosomes occur which
might be bivalent. Their number, however, is extremely small in this
case, and certainly does not include ring-shaped bivalents.

After diakinesis follows a contraction of the set of chromosomes
and after this metaphase. Spindle-fibres were seen in s. few cases but
no typical metaphase was found.

After clustering together the chromosomes gradually disperse
until eventually they run like a band through the whole cell. This
dispersion takes place in many different ways. Often, as in fig. 22 d,
some of the chromosomes remain together, while the others disperse,
but equally often dispersion takes place evenly. In this stage, the
„anaphase' ' of the first meiotic division, the chromosomes are generally
easy to count. Often 2 or 3 chromosomes lie close together or a few
may cover each other but there are no typical bivalents among them.

This renders counting occasionally uncertain but still I counted
mostly 36 and only a few cases were doubtful (35 or 36).

So I consider in this clone also 36 to be the normal somatic number.
This anaphase is very frequent in preparations and consequently
appears to take more time than anaphase in normal clones.

Spindle fibres sometimes exist. At late anaphase counting becomes
more difficult (fig. 22e) ; the angle in the anaphase-figure I frequently
observed, it proceeds the formation of telophase groups.

At telophase the chromosomes are always divided into more than
two groups often of very different numbers. Formation of both three
and four groups occurs constantly. The latter case is shown in fig. 22f.
The most frequently occurring situation is: one large group, two

Fig. 22. Meiosis in asynaptic Hevea (KN 220).
a. Normal pachytene, b. Asynaptic pachytene, c. Diakinesis with univalents,
d. Anaphase-I. e. Late anaphase, f. Telophase-I. g. Interphase, h. Anaphase-
II. i. Telophase-II. 1500 X.

groups of average size and one small group. Each of the groups forms
a daughter-nucleus with one or more nucleoli (fig. 22g). At interphase
therefore, 3 or 4 nuclei of various sizes may be found in each pollen
mother-cell. Bi-nuclear interphases I have never found, nor any
case of all the chromosomes being combined into one restitution-
nucleus.

At interphase the chromosomes are separately visible again and
can sometimes easely be counted. The interphase of fig. 22g shows
altogether 37 chromosomes. I suppose that one of the univalents has

Fig. 23. „Tetradsquot;, a. Normal tetrads, b. „Hexadsquot; and c. „octadsquot;
from asynaptic KN 220. 650 X .

been split at anaphase. Some of the chromosomes are also strikingly
smaller than others.

Interphase is followed by the second division which, rather curious-
ly, progresses quite regularly, apart from an abnormal number of
spindles. The spindle-fibres of three or four spindles are mostly clear-
ly visible. In fig. 22 h a pollen mother-cell with 4 anaphases of the
second division is shown. The chromosomes have become too small
to be counted at this stage, but it is easy to determine that even
groups of 2 chromosomes still enter into anaphase, and presumably
even sepeirate chromosomes which occasionally are found.

The number of grand-daughter-nuclei which is finally produced
depends upon the number of telophase groups produced by the first

division. If there are 3 groups, 6 nuclei are formed whilst from 4
telophase groups 8 grand-daughter-nuclei are produced.

These nuclei envelope themselves with a quantity of cytoplasm
which may correspond nearly with the number of chromosomes
included, thus instead of normal tetrads, „hexadsquot; and „octadsquot;
are formed (fig. 23b and c).

With KN 220 I have never found normal tetrads.
Young pollen-grains may be produced but they always degenerate
together with the stamens.

Before subjecting the species of a certain genus and various forms
of a certain sjjecies to a cytological investigation the identity of the
material to be investigated should first be carefully ascertained.

Hevea occupies in this respect a peculiar position. On the one hand
the plant has become very well-known owing to the enormous culture
plantations of H. brasiliensis, on the other hand its classification,
as I shall try to show, rests partly on a very insecure foundation.

There are a few species, it is true, that show very marked dis-
tinctions from others, as H. guianensis Aubl. and H. Spruceana
Muell. Arg., but there are also „speciesquot; which might as well be
varieties of another species, as products of hybridization.

For reasons stated above, I have used Ducke's classification; but
I think, with reference to the description of the supposed species-
hybrid H. collina X brasiliensis (Ch. 1, §2), some criticism to be
justified. Of the 5 groups into which Ducke (1930) divides the genus
Hevea the 2nd, 3rd and 4th are in the first place characterized by a
number of stamens which varies from 5 to 10 in incomplete or rudi-
mentary whorls. Also the species-hybrid H. collina X brasiliensis
has, as described in Ch. 1, § 2, flowers with a number of stamens
varying from 5 to 10.

Though the origin of this plant by interspecific hybridization is
not entirely certain, the probability of this origin is very great, and
in this case I think it is also very probable that species of Hevea that
are characterized by an irregular number of stamens have arisen by
hybridization of other species.

It will not be easy to supply conclusive evidence on this point but
crossing experiments with a number of species of Hevea would be
very helpful in testing the truth of my supposition.

Besides Ducke himself is also of opinion that natural species-
hybrids of Hevea actually occur, which is apparent from the hybrid
between H. Spruceana x hrasiliensis, sent by him to the Utrecht
Herbarium.

Bangham (1931) quotes a remark by T.F.C. (anonymus) (1920)
to the effect that: „Experience has shown that cross fertilization
between H. confusa and H. brasiliensis readily takes placequot;.

Bangham further is of opinion that the fact that all the species
investigated by him possess the same number of chromosomes:
„would suggest that fertile hybrids might be formed in some casesquot;.

In the opinion of these investigators, therefore, interspecific hybri-
dization of Hevea occurs.

The variation of the characters of a species like H. brasiliensis is
very great ; that the same holds good for other species appears from
the following remarks by Ducke (1933 p. 50):

„... de nombreux échantillons florifères et fructifères provenant
de plus d'un arbre et de plus d'une localité, pour chacune des espèces
plus fréquentes du Rio Solimoes et Rio Negro {guianensis, lutea,
Benthamiana, brasiliensis, membranacea et Spruceana) permettent
de connaître la variabilité très forte de ces espèces, même dans des
caractères que l'on avait jugés suffisants pour établir des sections du
genre...quot;

As a consequence Ducke, rightly in my opinion, considers Hevea
collina Huber a variety of H. guianensis Avë^. (Ducke, 1930, 1933),
but at the same time this again justifies doubt as to the existence of
a number of Hevea species, as the differences between the species
are often not much more considerable than those which appear
within one species like H. hrasiliensis.

The systematic classification of the genus Hevea, therefore, must
be regarded as uncertain.

The general progress of normal meiosis is, as is to be expected,
exactly the same with the various species.

Although fixations with Carnoy 3 : 1 and staining with Heiden-
hain's haematoxylin were only moderately successful, still the study
of meiosis has not resulted merely in a determination of the number
of chromosomes. The staining does not allow of a precise analysis of
early prophase; but that the number of chiasmata decreases from
diplotene till diakinesis, owing to terminalization and resolution, I

think may be accepted with certainty. At early diplotene I observed
4 chiasmata, but that bivalents with more than 4 chiasmata (5 and
6) occur is probable in view of the appearance of some already con-
tracting bivalents at middle diplotene (fig. 11).

At diakinesis I could not find more than 2 chiasmata and even
then in only 4 or 5 out of 18 bivalents. The remainder have only 1
chiasma, the situation of which varies from median interstitial (cross-
shape) to quite terminal (V-shape). Now it is a question whether
these interstitial chiasmata shift to the ends before metaphase.

The steadily progressing contraction of the chromosomes would
tempt to this assumption. The same applies to the bivalents which
have 2 chiasmata at middle diakinesis, one of them being terminal
the other subterminal. The differentiation in structure of the biva-
lents after diakinesis, however, is not great enough in the available
material to allow of an accurate verification, but when a bivalent
at late diakinesis is still cross-shaped further terminalization is not
probable.

In accordance with the behaviour of chromosomes and chiasmata
from diplotene till metaphase, 2 groups of plants may, roughly
speaking, be distinguished. In the first group little if any terminali-
zation takes place, as is e.g. the case with Fritillaria (Darlington
1930) ; in the second group terminalization is complete as shown in
Primula sinensis (Darlington 1931).

There are other forms whose behaviour is intermediary between
the two types mentioned (Moffett 1933).

It may be stated with certainty that Hevea does not belong to the
first group, for bivalents with interstitial chiasmata as found at meta-
phase in Fritillaria, decidedly do not accur in Hevea. That terminali-
zation in all bivalents of Hevea is complete may, however, be doubted.
Presumably Hevea will have to be included in the intermediary type.

Although for the object of a closer study of the chromosomes
separately, the number n = 18 is fairly large, and the chromosomes
themselves are fairly small, differences in size do exist. This is clearly
to be seen in the somatic metaphase-plates shown in fig. 2. The
largest chromosomes are at least twice as long as the smallest.

Added to this, „attachment constrictionsquot; are clearly visible in a
number of the chromosomes fixed with La Cour's 2BE, while they
do not occur in the same places in the different chromosomes.

In this investigation only a few somatic metaphase-plates were
inspected but an extensive investigation will perhaps render it possible
to group the chromosomes and to characterize them by means of
the size and place of constrictions.

The differences in size are of such a nature that at meiosis the
numbers of chiasmata wdll vary, because „the mean number of
chiasmata per bivalent is approximately proportional to the length
of the paired chromosomesquot; (Darlington 1932).

The present cytological investigation of Hevea has not yet opened
up new aspects with regard to taxonomy, as the species investigated
so far have all the same number of chromosomes. Bangham (1931)
already stated this but the number n = 17 mentioned by him is
wrong. Both in the various clones of H. brasiliensis and in the species
Spruceana and guianensis (besides the variety collina) I found n = 18
during meiosis in the pollen mother-cells.

Countings were done at diakinesis, metaphase-I, anaphase-I, mter-
phase and metaphase-II, leaving no room for doubt as more than 100
countings have been made. Moreover I found in root tips of material
from Surinam and of Sumatra seedlings 2n = 36.

Though the number of species investigated is still limited the
number 18 is found in all cases, so that 18 might be taken as the
„basic numberquot; of Hevea. 18 seems a fairly high basic number. Only
a few plants have a similar basic number, e.g. 19 in Salix and 17 in
Pyrus.

With a number like 18, however, one is inclined to think of tetra-
ploidy or triploidy, but a Hevea with 9 or 12 chromosomes has not
yet been found. Of other Euphorbiaceae, Euphorbia is the only one
which has had a great number of species investigated (Harrison
1930). Here the numbers 6 and 9 fairly often occur but other
numbers have been found as well. There are nowhere indications that
18 could be a secondary basic number for Hevea.

Because Heusser (1919) in his figures of various meiotic stages
persistently drew 8 chromosomes the thought has occurred to me
that Heusser may have been working with a form with 9 chromo-
somes. The absence, so far, of deviations from the fixed number 18,
however, renders this unlikely.

Bruun (1931) has succeeded in classifying 161 species of Primula
on the ground of morphological characteristics of the chromosomes.

„Average size and shapequot; are the principal characters and
further account has been taken of „the size and number of the chromo-
somes in relation to each other, basic number and appearance of the
constrictionsquot;.

The groups coincide with the sections, which facts prove the taxo-
nomic value of karyology. This might be of importance for Hevea.
A comparative karyological study of a great number of species might
open new perspectives for taxonomy.

Of much importance may be the occurrence of abnormal meiosis
in some male sterile and completely sterile clones of Hevea brasiliensis.

Typical male sterility in Viola Orphanidis has been described by
Clausen (1930); „Pollen sterile plants preferably are segregated
from plants with somewhat irregular meiosis, with either monosomic,
trisomic of tetrasomic behaviour of the chromosomes and the sterility
of the pollen may be caused by absence of a certain chromosome,
lost by trisomic or tetrasomic distributionquot;. Pollen sterility, however,
was also found with plants showing the normal chromosome number.

Somewhat abnormal meiotic behaviour is also reported by Winge
(1924) for speltoid aberrants of Triticum and by Huskins for fatuoid
Avena (1927) and for speltoid Triticum (1928). Here univalents and
trivalents occur constantly and especially in hybrids lacking a
chromosome, but are also found in hybrids showing the normal
chromosome number with partly abnormal combinations.

Typical male sterility, however, is not found where both sexes are
sterile. None of the above mentioned cases agrees completely with
the pollensterile Hevea clones. In the latter irregular meiosis with
the occurrence of uni- and multivalents always is followed by
degeneration of the cells. This happens after formation of tetrads or
uninucleate pollen-grains.

With V. Orphanidis the pollen mother-cells degenerate during the
first meiotic prophase as a result of an incompatible set of chromo-
somes originating in meiotic irregularities in the previous generation.
In fatuoid Avena and speltoid Triticum some meiotic irregularities
occur but without typical male sterility.

When discussing whether pollen sterility in Hevea may be caused
by the lack of a chromosome or by gene action, the second expla-
nation seems to be more probable than the first, as in interkinesis
2 x 18 chromosomes are found most frequently.

Appearance of univalents and multivalents often is connected
with polyploidy. An extreme condition in this respect is shown e.g.
by Meurman (1929) in Prunus Laurocerasus. We do not know
anything about polyploidy in Hevea but the occurence of multi-
valents might be an indication of it.

About the asynaptic clone KN 220 a few things may be said.
Asynapsis has been found in several plants. Well known are the
investigations of Rosenberg (1917, 1927) on the semi-heterotypic
division in parthenogenetic Hieracium. In the meiosis of the „Leviga-
tum typequot; only univalents occur, in the „Boreale typequot; univalents
and a varying number of bivalents.

Ljungdahl (1922) reported total asynapsis in hybrids of Papaver
ailanticum X dubium, Karpechenko (1924, 1927, 1928) in hybrids
of Raphanus sativus x Brassica oleracea, Goodspeed and Clausen
(1927) in Nicotiana Bigelovii X glutinosa. Partial asynapsis occurs
frequently in species hybrids; a number of such cases are given by
Darlington (1932). Further asynapsis has been described for dwarf
oats and wheat by Huskins (1927) and Huskins and Hearne (1933)
and for Primula kewensis by Newton and Pellew (1929).

In the above mentioned cases of asynapsis, the direct cause is not
always the same. In the Levigatum type of Hieracium asynapsis may
be caused by hybrid nature. In the cases of Raphanus X Brassica,
Papaver and Nicotiana the hybrids show the same somatic chromo-
some-number of the parents, but the chromosomes fail to pair, which
may be due to low affinity. When the parents possess an unequal
number of chromosomes, the hybrids show mostly bivalents and
the rest univalents.

Total, partial and varying „asynapsisquot; is sometimes caused by
the lack of a chromosome as is the case with dwarf oats. Primula
kewensis and occasionally in poUen mother-cells of Viola Orphanidis
(Clausen, 1930).

That the parents of KN 220 would have belonged to different
species, must be excluded; consequently asynapsis as a direct result
of species hybridization is out of the question.

Asynapsis might further be due to the lack of a chromosome; in
KN 220 I found mostly 36 chromosomes in the anaphase-I and once
37 in the interkinesis; so the lack of a chromosome is not probable.

Finally asynapsis may be caused by the action of an „asynaptic
genequot; (Beadle 1933) but on this possibihty no information as to
Hevea can be obtained.

The course, as well as the cause, of asynaptic meiosis differs in many
plants. A different course may be taken in different pollen mother-
cells of the same plant. In Viola Orphanidis it is almost normal,
apart from the occurrence of quadrivalents.

With the Levigatum type of Hieracium diakinesis is followed by
interkinesis with the omission of the first meiotic division, this
leading to the formation of dyads and of gametes with an abnormally
high chromosome number. A similar process is found in Raphanus-
Brassica hybrids, in hybrids of Nicotiana tabacuni x Rusbyi (Brie-
Ger, 1928) and in dwarf oats.

With KN 220 I never observed formation of dyads. The irregular
first division is always followed by a regular second division and
always resuhs in the formation of „hexadsquot; and „octadsquot;.

I have never met a similar constantly occurring deviation in the
literature on this subject.

Besides dyads and tetrads, groups consisting of three to' seven
cells are formed in Raphanus-Brassica hybrids; in Musa (Tisch-
ler 1910 and Cheesman 1932) „tetradsquot; are formed which also
contain varying numbers of cells, usually more than four. In such
forms which also show irregular divisions. White (1928) found
univalents.

Much resemblance with KN 220 exists in hybrids of Nicotiana
Bigelovii x glutinosa. With the latter, however, formation of abnor-
mal tetrads takes place, especially as a result of irregular anaphase-II,
whereas in KN 220 it is due to irregular anaphase-I.

As is shown by Beadle (1933) in asynaptic maize, the existence
of unpaired chromosomes in metaphase, diakinesis and even in
diplotene does not always mean complete failure of pairing in zygo-
tene. Synapsis may be followed by separation during pachytene
without chiasma formation.

So the lack of a typical pachytene in our ,,asynapticquot; Hevea does not
exclude the possibility of zygotene pairing. With many plants it is
possible to check the cytological behaviour of the chromosomes by
genetical experiments on crossing-over. With Hevea, however, this
is practically impossible as the offspring of a tree, when the latter is
artificially pollinated, require five years before the next cross or
selfing can be made.

The cytology of Hevea will be thrown on its own as is the case with
many cultivated plants.

From the present investigation it is evident that the chromosomes
of Hevea are worthy of a further stady both in normal meiosis and
irregular meiosis as well as in somatic divisions. So I hope to be able
to continue the investigations on this subject.

gmanensis Aubl. H. collina Huber. and of two species-hybrids' H
spruceana x brasiliensis and H. collina x brasiliensis.

In connection with the description of the species-hybrids and
with the great variation of characteristics in Hevea the value of the
recent taxonomy of Hevea is discussed.

2 As far as possible the behaviour of the chromosomes in meiosis
has been investigated. From early diplotene to diakinesis terminahs-
ation of chiasmata and a decrease of the number of chiasmata from
2 6 to 1—2 takes place. Terminahsation is not entirely complete
the one or two chiasmata in late diakinesis being terminal or inter-
stitial and even central (cross-shape).

3.nbsp;The haploid number of chromosomes has been proved to be 18
m the pollen mother-cells of all the species and forms investigated
Countings were made in diakinesis, metaphase-I, anaphase-I inter-
phase and metaphase-II, the total number amounting to more than
hundred meiotic stages, leaving no room for any doubt

In metaphase plates of somatic divisions in root tips 36 chromoso-
mes were found.

4.nbsp;In some cultivated clontsoi Hevea brasiliensis showing partial
or total male sterility, irregular meiosis takes place which is coupled
with the occurrence of univalents and muhivalents; of the latter
trivalents are most common. The same irregularities were found in
the hybrid of H. Spruceana x brasiliensis.

5.nbsp;A case of asynapsis in a clone of H. brasiliensis, showing com-
plete female and male sterility, has been described and compared
with asynaptic meiosis in other plants. There is an almost complete
lack of chromosome pairing in the prophase. A typical metaphase-I
has not been found, the chromosomes after prophase being scattered
in the cytoplasm. Three or four daughter-nuclei of different size are
always formed, the chromosomes of which, curiously enough, show
a very regular second division, resulting in the formation of he'xadsquot;
and „octadsquot;.

Bangham, W. N. (1931). Chromosomes of some i^ewa species. Journ. Arnold
Arbor. 12, 287.

Bëlar, K. (1923). Die cytologischen Grundlagen der Vererbung. Handbuch
der Vererbungswiss. Berlin I, 1.

Bremer, G, ( 1923). A cytological investigation of some species and species-
hybrids within the genus Saccharum. Genetica 5, 97.

Bruun, H. G. (1932). Cytological Studies in Primula. Symbolae Botanicae
Upsaliensis T, 1.

Cheesman, E. E. (1932). Genetic and Cytological Studies of Musa I. Journ.
Gen. 26, 291.

Clausen, J. (1929). Chromosome Number and Relationship of some North
American Species of Viola. Ann. Bot. 43, 741.

Darlington, C. D. (1930). Chromosome Studies in Fritillaria III. Cyto-
logia 2, 37.

Darlington, C. D. (1931). Meiosis in diploid and tetraploid Primula sinen-
sis. Journ. Gen. 24, 65.

Ducke, A. (1930). Plantes nouvelles ou peu connues de la région amazo-
nienne IV. Arch. d. Jard. Bot. d. Rio de Janeiro, vol. 5, 99.

Frey-Wyssling, A., Heusser, C. en Ostendorf, W. (1932). Het identifi-
ceeren van Hevea-cloonen op jeugdigen leeftijd. Arch. v. d. Rubberc. 16, 1.

Geerts, t. h. (1909). Beiträge zur Kenntnis der Zytologie und der partiel-
len Sterilität von Oenothera Lamarckiana. Recueil d. trav. bot. néerl. 5, 93.

Heusser, C. (1919). Over de voortplantingsorganen van Hevea brasiliensis
Müll. Arg. Arch. v. d. Rubberc. 3, 455.

Hsu Chuan Tuan. (1930). Picric acid as a destaining agent for iron alum
hematoxylin. Stain Technology 5, 135.

Huber, j. (1905). Ensaio d'uma Synopse das Especies do genero Hevea.
Bol. d. Museu Goeldi 4, 620.

Huskins, C. L. (1927). On the Genetics and Cytology of Fatuoid or False
Wild Oats. Journ. Gen. 18, 315.

De flora van het Middellandsezeegebied staat in nauw
historisch-plantengeographisch verband met die van Zuid-Afrika.

F. Markgraf, Genetische Beziehungen der Mittelmeer-
flora. Ber. d. Dtsch. Bot. Ges. 52, 68.

Het is wenselijk dat van academici, die zich op het gebied
van het onderwijs willen begeven, praktische oefening in het
lesgeven, benevens eenig inzicht in de psychologie van de rijpere
jeugd en in de didakdek, geeist wordt.