[t
f)

•■.X I. y--. -îs

.'^'lA':'-

■y,'- -y

gt; ti* -V.

PROEFSCHRIFT TER VERKRIJGING VAN DEN
GRAAD VAN DOCTOR IN DE WIS- EN NATUUR-
KUNDE AAN DE RIJKS UNIVERSITEIT TE
UTRECHT OP GEZAG VAN DEN RECTOR MAG-
NIFICUS DR. C. W. VOLLGRAFF, HOOGLEE-
RAAR IN DE FACULTEIT DER LETTEREN
EN WIJSBEGEERTE, VOLGENS BESLUIT VAN
DEN SENAAT DER UNIVERSITEIT TEGEN DE
BEDENKINGEN VAN DE FACULTEIT DER WIS-
EN NATUURKUNDE TE VERDEDIGEN OP
MAANDAG 9 DECEMBER 1935, DES NAMIDDAGS
TE 3 UUR

'S-GRAVENHAGE

MARTINUS NIJHOFF
1935

BIBLIOTHEEK DER
RUKSUNIVERSITEIT
UTRECHT.

' • t quot;V

Het zij mij vergund, hier in eerbiedige herinnering vóór alles te ge-
tuigen van de dankbaarheid, die ik gevoel jegens wijlen Prof. Dr. F.
A. F. C. Went, den leermeester, wiens belangstelling in mijn werk en
leven mij vele malen tot steun is geweest.

Met groote waardeering en erkentelijkheid gedenk ik het ouderlijk
huis, waar aanleg en omgeving samenwerkten, om liefde voor de
natuur te ontwikkelen en haar schoonheid te leeren genieten.

Met bewondering en eerbied denk ik terug aan het onderwijs van
wijlen Dr. J. J. le Roy, die niet slechts op degelijke en boeiende wijze
de ^^levensverschijnselen van den menschquot; behandelde, maar wiens
wijsgeerige geest daarbij belangstelling wist te wekken voor den „ont-
wikkelingsgang van het menschelijk denkenquot;.

Zeergeleerde Hoogenraad, ongetwijfeld hebt Gij indertijd den
grootsten invloed op mij uitgeoefend. Niet slechts door Uw bekwaam-
heden als docent, maar evenzeer door Uw gaven als mensch, zijt Gij
in staat de lust tot eigen waarneming en onderzoek te wekken en te
leiden bij Uw leerlingen, zonder dat zij daarbij eigen aard behoeven
te verloochenen. Voor Uw voortdurende belangstelling en vriend-
schap, die ik reeds in zoo ruime mate genoten heb, zal ik steeds dank-
baar blijven.

Hooggeleerde Jordan, Koningsberger, Nierstrasz, Pulle en
Rutten, zeer erkentelijk ben ik voor wat ik van U mocht leeren of
voor de welwillendheid, die Gij mij wel hebt betoond. Hooggeleerde
Pulle, ten zeerste waardeer ik het, dat Gij steeds bereid zijt geweest
mij met Uw raad bij te staan en dat Gij mij bij het bewerken van dit
proefschrift in enkele moeilijkheden van advies hebt willen dienen.

Aan Uw practica in de Protistologie, zeergeleerde Entz, bewaar ik
nog de meest aangename herinneringen.

ik U dank verschuldigd. Dat Gij mij op zoo onbekrompen wijze mate-
riaal en hulpmiddelen verschaft hebt, om dit onderzoek te kunnen
uitvoeren, waardeer ik meer dan ik zeggen kan; dat Gij, ondanks Uw
drukke bezigheden, steeds tijd hebt gevonden voor kritische bespre-
kingen, die aan het werk ten goede zijn gekomen, stel ik op hoogen
prijs. De uren in Uw gastvrij laboratorium doorgebracht, zullen voor
mij dan ook van groote beteekenis blijven!

Waarde Berrington, wees verzekerd, dat Uw hulp bij de correctie
van de Engelsche tekst ten zeerste door mij wordt gewaardeerd; ook
voor Uw medewerking, waarde vriend Reilingh, ben ik nog zeer
erkentelijk.

Ook U, zeergeleerde Gerrits, ben ik erkentelijk voor de wel-
willendheid, mij betoond bij de regeling der lessen en voor Uw
belangstelling in dit werk.

Zeergeleerde Mej. Polak, gaarne richt ik ook tot U een woord van
dank voor de vriendelijke wijze, waarop Gij mij steeds de literatuur
hebt willen verschaffen, die ik noodig had.

Waarde Jansen, Knoop en van de Peppel; Gij hebt mij vele
malen door Uw hulp en voorkomendheid het werken in tuin en labo-
ratorium vergemakkelijkt; daarvoor zeg ik U hartelijk dank.

Had ik niet de voortdurende vaardige hulp gehad van mijn vrouw
bij de technische bewerking van het materiaal, dan was het nu ver-
kregen resultaat nog niet bereikt; haar blijvende belangstelling in
den voortgang van het onderzoek heeft mij de uren van gemeen-
schappelijken arbeid onvergetelijk gemaakt.

a.nbsp;to describe the general course of the meiosis of the pollen mother
cells,

c.nbsp;to trace how far with the representatives of the genus Canna
particularities appear in the meiotic phenomena which are of im-
portance for the contemplation of the method of pairing of the
chromosomes.

Through former cytological researches by Honing (1928) the expec-
tation was roused that in this material something might be found for
a critical inspection of the problem about para- and telosyndesis.

The cytological literature with regard to species and varieties of
the genus Canna is not very extensive, most documents about that
matter are of a systematical and morphological nature, especially,
in connection with the remarkable structure of the flower (Costerus,
1916, 1920). After 1900, also this genus was used for genetical re-
searches ; for several forms the number of chromosomes was examined,
but so far attempts to make a connection between cytological
and genetical occurrences have not been made (apart from a single
exception, Belling, 1921, 1925, Honing, 1915, 1923, 1928).

The material used for the research written down here was ob-
tained from the Canna-cultures of Prof. Honing, who for his genet-
ical researches grows a number of forms in the greenhouses of the
„Laboratory for Genetics of the University College of
Agriculturequot;, Wageningen. I would tender him my best thanks
for his kind permission to collect material in his experimental garden.

world; the cytologically worked material was collected from genetic-
ally well-known plants.

Cannaceae I, 1912) a historical review of the research of the genus
Canna, which is worth studying for everybody who is occupied with
these plants.

Very likely Cannas were already grown in Europe at the end of the
16th century. The first clear statement about that occurs in Bau-
hin's Pinax Theatri Botanici which was printed in 1623. These plants
are also mentioned by Tournefort, Piso, Rheede and Rumphius.

For the first time three species are mentioned in Linne's „Species
Plantarumquot; viz. Canna angustifoUa, Canna glauca and Canna indica\
the latter with a diagnosis of little significance and with a collection
of varieties representing almost as many species, so that Kranzlin
remarks: „Vom allerersten Anfang an ein Stein des Anstossesquot;.

In the 19th century the interest for Canna is evidently ex-
ceptionally great, as well from the side of scientific researchers as
from the side of the practical growers.

Many forms were collected and described in the middle of last
century, of which the herbar-materials are deposited in Berlin—
Dahlem at present. These collections are worked by Kranzlin and
reduced by him to 51 species: 43 species from America, 6 from Asia,
1 from Africa and 1 from unknown origin.

Suppose we take this number of species as a basis and state that
under the influence of the interest of the horticultural artists,
hybrids have been grown and again hybrids of hybrids, then it is
possible to understand that from cultivated forms which at present
are scattered all over the tropical and subtropical parts of the world
as „garden-plantsquot;, the exact systematical place cannot always be
directly and easily determined. For it must be added that many
growers controlled insufficiently or not at all what had resulted
from the crossing; the statements of the names of crossed species are
often altogether wrong (sometimes even purposely(?) misleading,
according to Costerus, 1916a). It is quite well possible that so-called
„wildquot; forms are escaped garden-plants and of hybrid origin. Un-
doubtedly forms of a very different origin, bearing the same name,
have been in cultivation, and so it may be explained that different
statements are to be found for the „samequot; kind of plants, with regard
to the number of chromosomes. As an example: for Canna indica are
to be found for the haploid number the statements 3, 8, 9; for the
diploid number 9 9/2, 27/2 (Gaiser, 1926, 1930«, 19306; Tischler,

Some authors mention Canna sp., Canna hybrids, „garden-va-
rietiesquot;, without further statements, but that is why it is not always
possible to compare precisely their results.

As already pointed out the present classification of the species in
the genus Canna is due to the research of Kränzlin (1912) and later
writers have adopted his diagnoses and tables (Winkler, 1930;
Wettstein, 1935).

In order to make comparison of examined plants possible, it is
now desirable to use Kränzlin's tables as a basis for the description
and the denomination of the species.

The forms, cytologically described (in this investigation) belong
to the subgenus Eucanna, which is characterised by the presence of
two or three petal-like staminodes in addition to the labellum. This
subgenus is divided into :

Sectio I, Bialatae (with 2 staminodes) and
Sectio II, Trialatae (with 3 staminodes).

Canna humilis Bouché (Kränzlin, 1912, p. 43) received from the
botanical gardens at Montevideo, is a species with two staminodes
and with totally green and glabrous leaves. The plants used slightly
differ from the description given by Kränzlin, but yet in my opinion
this form must be ranked among the species. The variability of the
characteristics of cultivated forms is apparently very considerable,
especially as regards the quantitative properties (cf. Honing, 1923,
p. 4). Kränzlin (I.e. p. 12) says:

„Die Variabilität der Färbung bei Exemplaren derselben Art in
Regeln zu fassen, muss ich für ein zunächst wenigstens aussichtsloses
Unterfangen haltenquot;. The leaf-blades of Canna humilis are light-
green, white-margined, ovate. (70 cm. x 23 cm.) The bracts are light-
green with pinkish margin. (±30 mm. long, ±15 mm. br.) Ovary:
green with red. Sepals : crimson to brownish-red, the middle greenish,
the outer surface glaucous. Petals: carmine, upwards light-red. (40 mm.
long.) The inner surface yellow-green, striped. Labellum: the underside
of the curved part orange with red margin, the upper side is yellow-
orange and red-spotted, the apex is red. Staminodes: two, bilobate,

Canna discolor Lindl. (Kranzlin, 1912, p. 38). The variety exam-
ined is also a form, which does not altogether correspond with the com-
pared descriptions. Kranzlin attaches great value to the number of
petal-like staminodes, for he uses it as a basis for the classification of
the subgenus ZiMcawM«; according to this classification the mentioned
variety must be considered to belong to the Bialatae, although very
seldom (twice out of some thousands of flowers) a third staminode
has been observed (by Honing) noticeably but not very much smaller
than the outer ones. The exceptionally great development of the
dark-red colour in stems, leaf-sheaths, leaf-margins, bracts, inflor-
escences, flowers and ovaries, is the most prominent of all the
characteristics of this variety. Leaves: ovate, totally glabrous. Leaf
blades: dark-green, shiny, with dark-red margin and middle-vein,
the top very acute. All parts of the inflorescences are totally dark-red.
Bracts: glaucous (25 mm.L, 20 mm. br.) Sepals: dark-red, (15—20
mm. 1., narrow). Petals: linear lanceolate, the outher surface deep-
red, the inner side slightly greenish, (35—40 mm. 1.) Staminodes:
generally 2, bilobate, deep-red. Labellum: red, the middle flesh-
coloured, red-spotted (fig. 1 and fig. 2). It is, of course, quite well
possible that the characteristics of this variety, which differ from
those of the pure Canna discolor, are due to a previous crossing with a
trialate species (such as Canna indica L).

Canna lutea Mill (Kranzlin, 1912, p. 39) also received from the
botanical gardens at Montevideo, is a bialate species with light green
leaves, the leaf-sheaths and leaf-blades are glabrous (the form is that
of Canna discolor). Bracts: light-green, (30 mm. 1.), glaucous. Ovary:
light-green. Sepals: light-green (14 mm. 1.), glaucous. Petals: light-
yellow, towards the top greenish, (45 mm. 1.). Labellum: light-yellow,
upper side red-spotted. Staminodes: 2, the same colour as the label-
lum, both sides with light-red spots and stripes; top bilobate or
sometimes nearly entire (The appearance of the flower may be judged
from the photograph, fig. 1). The above mentioned Cannas (C. hu-
milis, discolor and lutea) are, no doubt, closely related.

Canna glauca L. (with the varieties ,Javaquot;, „Boliviaquot;, „Mon-
tevideoquot; and „Pure yellowquot;), a species of the Trialatae.

For the description of these forms we may refer to the publications
by Honing and to the photograph, fig. 1, showing the form of the

flowers of the varieties examined. It was not possible to identify
these forms more accurately with forms given by Kranzlin; there-
fore the names mentioned by Honing, are used.

A historical review of the literature about the cytology of the
reproduction-organs can be relatively short here, because in the
preceding century only little or nothing at all was written with regard
to the research of the chromosomes in species of the genus Canna
(cf. ScHijRHOFF, 1926; Schnarf, 1931).

Before 1900 the cytological publications chiefly relate to the
embryogenesis in general, but they do not yet contain information
about the origin and the division of the chromosomes.

The first essay in which something is mentioned about the
number of chromosomes and the reduction-division of Canna indica,
appeared in 1900. It is a publication by Karl M. Wiegand: „The
development of the embryosac in some monocotyledonous plantsquot;.
He describes and compares the development of the embryosac in
Convallaria majalis L., Potamogeton foUosus Raf. and Canna indica L.
His conclusion is: „The development in Canna was found to be
nearly normalquot;. In the summary of his paper he says with regard
to Canna:

„The number of chromosomes in the vegetative divisions is six.
When passing to the poles at the heterotypic division there were still
six; but later the second division showed only three as the reduced
number. Probably the segmentations for both divisions occur during
the prophase of the heterotypic division. This number is one of the
smallest yet found in vegetable tissuequot;.

The illustrations belonging to this paper are clear and do not
suggest any doubt about the number of chromosomes. An anaphase-
representation clearly shows the separation of two sets of six chromo-
somes each. About the development of the chromosomes Wiegand
did not see much, which appears from his words: „The nuclei and
chromosomes in Canna are so small that little could be done toward
working out the segmentation of the latterquot;. Possibly he means this
as compared with the chromosomes of Convallaria and Potamogeton;
it is not to be stated with certainty because neither from the illus-
trations themselves does it appear nor is in the text the enlargement
of the figures mentioned. Though the prophase stages presented
difficulties in observation, this was not the case with the following
ones, for he says further on: „the ordinary process of division assures

six daughter chromosomes for each resulting nucleus. This count was
made many times with great ease, owing to the small number of
segments and always with the same resultquot;.

Still his statement must be assumed with some reservation for
he also says that he found more than six chromosomes in the arche-
sporial nucleus. Moreover it must be taken into consideration that
the number of countings of the material was not large: „Only two
spindles were found representing the heterotypic division and these
both showed the globular daughter segments on their way to the
poles (anaphase). The four counts here made gave in every case the
number of six, instead of three as one would expect after reductionquot;.
And he only saw one or two sections with the second division.

It is not to be settled whether Wiegand dealt with a deviating
form of Canna indica or with abnormal material; anyhow, the obser-
vations mentioned by him were not affirmed later on. Soon after
the appearance of the discussed essay a doubt was expressed about
the truth of the statements made by Wiegand.

In 1903 M. Koernicke gave a critical report about the state of the
cytology, the knowledge of protoplasma and nucleus: „Der heutige
Stand der pflanzlichen Zellforschungquot;. Koernicke in order to con-
trol the result of Wiegand, made the same investigation once more.
„Die eigentümliche Art und Weise, auf welche sich der Reduktions-
vorgang in den Embryosackmutterzellen von Canna nach Wiegand
vollziehen soll, veranlassten mich zu einer Prüfung der Angabenquot;.
„Die Resultate meiner Nachprüfung, welche ich hier anfügen möchte
waren folgende: Zunächst traten mir bei Beginn der ersten Teilung
acht längsgespaltene Chromosomen entgegen, die sich stark ver-
kürzten. Im Äquator der ersten Spindel finden wir die acht Doppel-
klümpchen wieder, deren Hälften nach den Polen auseinander-
weichen. Bei dem zweiten Teilungsschritt fand sich die Achtzahl
der Chromosomen wieder. Dieselbe Zahl trat mir auch in den beiden
entsprechenden Teilungsschritten der Pollenmutterzellen, die ich
daraufhin studierte entgegen. Nach den Beobachtungen die ich bei
Canna machte hegt für mich die Annahme nahe, dass Wiegand
Spindeln mit unvollkommen fixierten, verklumpten Chromatin-
elementen vor sich hatte, die mir auch hier und da entgegentraten.
Wass die niedrige Chromosomenzahl (6) anbetrifft, die Wiegand für
die vegetativen Kerne angibt, so muss ich auch die Richtigkeit dieser

Angabe in Frage ziehen. Bei allen meinen Zählungen konnte ich
immer bestimmt das Vorhandensein von mehr als zehn konstatie-
renquot;.

Koernicke does not give illustrations in his article; but, anyhow,
it is clear that he saw eight gemini in the development of embryosac
and pollen, whereas he saw more than ten chromosomes in the vege-
tative cells. Neither with this research was the number of chromo-
somes in Canna indica stated with absolute certainty.

Baur in the first edition of his book „Einführung in die ex-
perimentelle Vererbungslehrequot; (1911) takes Canna as an example
for the mechanical interpretation of Mendelian heredity under
recognition of its having three chromosomes in gametic cells.

Honing published in 1915 „Kreuzungsversuche mit Canna-
varietätenquot;. Obviously his intention was „zu prüfen, ob die Mendel-
spaltung auf der Verteilung der väterlichen und der mütterlichen
Chromosomen bei der Reduktionsteilung beruhen kann oder nicht.
Canna indica sollte nämlich generativ nur drei Chromosomen be-
sitzenquot;. The genetic research however proved that so small a number
was not very probable. That is why Honing refers to the statement
of Koernicke that Canna should have eight chromosomes in the
generative cells. Through countings in vegetative cells he found
sixteen as the diploid number. Later on (Canna crosses I, 1923)
Honing mentions that he could confirm for his strain of Canna indica
the number of chromosomes found by Belling as haploid 9.

At that time several investigators tried to obtain a more accurate
knowledge of the relation between chromosomes and the external
characters of an organism. Tokugawa and Kuwada published an
investigation in 1924 which was made already before 1919; it
is a study of diploid and triploid garden-varieties of Canna. The
diversity of the statements of Wiegand and Koernicke induced the
writers to re-investigate the number of chromosomes of Canna]
moreover they made a comparative study in somatic characters
between the diploid and triploid forms. (For the names of these
varieties, see: Tokugawa and Kuwada, 1924). The numbers of
chromosomes (investigated in the root-tips) are 18 and 27 (var.
H alley Comet only with 17—18) ; 6 or 16 could never be found. (They
gave 16 figures of somatic metaphase plates : 11 with 18 chromo-

somes and 5 with 27 chromosomes; somatic pairing or clear con-
strictions are not to be seen). „The number of chromosomes in the
meiotic nuclear division was examined in several garden varieties.
In Eldorado, Halley Comet and Black Warrior 9 gemini were counted
with accuracy, as was to be expected from the number of chromo-
somes in the root-tipsquot;. In the triploid varieties the expected
number of gemini and unpaired chromosomes was also found; the
authors give a figure of an anaphase-spindle with 27 single or un-
paired chromosomes scattered in the karyoplasm. „The process of
meiosis in Canna is generally of somewhat abnormal tendency and
in some cases it was almost impossible to count the exact number of
chromosomes. The synapsis takes place as usual, but the linin sub-
stance seems to be rather poor in quantity. In the next stages the
meagerness of the linin substance becomes more apparent. We can
find only lines of chromatin-beads, which remind us of nuclear
threads, but the connection between the beads is not so rigid as is
usually the case in the latter. In the diakinesis the chromosomes are
very loose in construction, sometimes so much so that one chromo-
some might easily be mistaken for two or even more. In such cases it
is almost impossible to count the exact number of chromosomes or
gemini. In Eldorado, Halley Comet and Black Warrior the behaviour
of the chromosomes was somewhat regular, which enabled the
writers to count the exact number as already mentioned. The division
does not always result in only two daughter nuclei, as is usually the
case, but very often three or more nuclei of various sizes can be
found and many chromosome-like bodies with a clear circular space
around each body. Thus we find more than one nucleus in a daughter
cell. In these cases the nuclei may often appear as a larger nucleus of
irregular form, connected with each other by a slender nuclear bridge
a figure which reminds us of amitosis. The homotypic division is not
yet fully studied, but can by no means be normal. So we have two
or even more nuclei in a pollengrain, and not infrequently also
chromosome-like bodies in the cytoplasm. The writers actually found
a pollen-grain having five nuclei. In case two or more nuclei are
found in a daughter-cell produced by the first division, they behave
quite independently of each other in the second division, making the
spindle in any direction as the chance may be. From this irregularity
a disorder in the arrangement of cells produced from a pollen mother

cell followsquot;. „The cell division is abnormal, especially in the second
division. In the first division we not infrequently find a newly formed
boundary wall appear in section, not as a straight hne as is usually
the case, but undulated. In the second division abnormalities occur
in a higher degree. In Canna, in the majority of cases the wall pro-
duced in the second division makes an acute angle to the first wall
instead of a right angle as is normally the case in Monocotyledons and
sometimes it appears even parallel to the latter. These abnormalities
seem to be, though to a different extent in the different varieties, the
usual case in Canna, either diploidal or triploidal. Thinking the ab-
normalities in the meiosis of the pollen mother cells might have
caused the sterility of the plants at least in part, the writers examined
the pollengrains of some garden varieties, and found in each case
a certain amount of sterile pollengrains. Some garden-varieties pro-
duce seeds. Only a few of them barely germinate, but most of them
do not come through. In the seeds examined we found most of the
embryos degenerated. This, perhaps, has some relation to the
abnormal combination of chromosomes due to the irregularities in
the meiosis of the pollen mother cellsquot;.

It is remarkable to state how little has been said with certainty
about the course of the prophase in former researches; it is probably
a consequence of the fact that the earlier stages of development are
difficult to investigate. Neither do the above authors say much about
this matter and only give one drawing of a synaptic contraction and
of a later stage of the var. ^^Meteorquot;. Probably it is a figure of early
diplotene; the large paired granules are evident but the connecting
threads are not shown in the picture. In all probability the authors
could not see the connections, which is not very astonishing for the
observation is difficult, because the achromatic substance of the
chromosomes in Canna is nearly invisible in some stages, so that the
chromatin-granules seem to Ue separately in the nucleus.

In 1921 John Selling's article appears: „The behaviour of ho-
mologous chromosomes in a triploid Cannaquot;. In the preceding years
much cytological work has been done and the interest is directed to
the pairing of the chromosomes in prophase and metaphase and to
the separation in the anaphase; it is to be expected that researches of
polyploid forms will give a better comprehension of these phenomena
about which there are still so many different opinions. Belling

says: „In 19201 grew 46 differently named clones and species of Canna,
obtained, without particular selection, from three leading dealers of
New-York, Philadelphia and Florida and from the U.S. Department
of Agriculture. The chromosomes of 31 of these were counted in the
first or second divisions of the pollen mother cells; about 250 groups
being drawn with the camera. Of these 31 clones, 22 showed nine
dyads or bivalents which split into 9 9 at the first division; 3 clones
had a total of 18 single chromosomes (or 9 bivalents), which divided
at the first division into 8-f 10, or other unequal numbers, not com-
monly into 9 9; 5 clones were probably completely or nearly

smaller number of chromosomes after the first division than the
triple number (in these cases 24 to 26 instead of 27); while one clone
was regularly triploid, showing nine triads at the prophase and first
metaphase, and a total of 27 chromosomes after the first division.
Thirty-two pollen mother cells were drawn with the camera. In 18
of these the total number of chromosomes could be accurately
counted, and was 27quot;. This regularly triploid clone was obtained
from Thornburn, New-York in 1920 under the name „Gladiatorquot;.
It is a smal form as the diploid Cannas. (In the paper of Tokugawa
and Kuwada, „Gladiatorquot; is a diploid form! The investigators gave
no further information in the text neither pictures).

In his summary Belling states: a) „Most of the Cannas examined
were diploid, showing nine dyads, before the first division in the
pollen-mother-cells, and these in most plants separated into 9 9.
V) One of the triploid Cannas showed commonly nine triads, each of
which separated into two and one on the spindle, in a random manner
with regard to the two polesquot;.

The already mentioned researches of Tokugawa and Kuwada of
1924 have been published in consequence of Belling's paper of 1921.

In 1925 Belling publishes again about CawM«. (Belling 1925a and
19256). In the first paper he mentions Canna only as an example of a
genus with triploid forms and he observes, that „the chief character-
istic of true triploids is a partial or total sterility; the triploid
Cannas, for instance, never (or rarely) producing seedsquot;. „The
characteristic of the tetraploids is the possession of much larger
cellsquot;. The production of a triploid apparently originates with the
doubling of the chromosome number in one of the gametes. This

doubling is in many cases certainly due to a temporary chill, resulting
in non-reduction or non-division during the maturation divisions of
pollen mother cells, or megaspore mother cells (or to non-division in
the somatic mitoses which give rise to the sub-epidermal tissue of a
branch). Triploid plants are already of commercial value for their
flowers in the case of a few Cannasquot;. Belling says further, that the
production of triploids should be tried by exposing the young flower
buds to low temperatures.

The second article gives the figures, belonging to Selling's
publication of 1921. The preparations were made with the iron-
aceto -carmine method. The pictures show:

a) Metaphase I in a pollen mother cell of the ornamental Canna
^filadiatorquot; with 9 trivalents. The connections between the chromo-
somes of each trivalent are different; they can form a chain, a V or a
tri-radial. Their finer structure is not visible in the pictures.

h) The same stage of the Burpee Canna „Firebirdquot;, which clearly
shows that 27 chromosomes form 9 trivalents.

c) The reduction division in the Canna ^^Pennsylvaniaquot;, which
shows trivalents, bivalents and univalents.

Meanwhile the aceto -carmine method was put into practice more
and more for preliminary investigations, especially for countings of
chromosomes. Heitz publishes in 1926 an article about the appli-
cation of this method, in consequence of the results of an investi-
gation. „Ich untersuchte Canna flaccida und zwei Gart en Varietäten
von C. indica, ebenfalls Wurzelspitzen. Für die eine könnte gleich
im ersten Präparate die Zahl 18 mit Sicherheit festgestellt werden,
für die zweite mit groszer Wahrscheinlichkeit. Es liess sich nicht
entscheiden, ob 2n=18 oder 20 ist. Dagegen zeigten die Platten der
noch nicht untersuchten Art C. flaccida einwandfrei ebenfalls 18
Chromosomenquot;.

Hüning gives in 1928 a review of his new investigations of pollen-
mother cell divisions. He determined the chromosome number of
Canna glauca and aureo-viltata, which is haploid 9. For the greater
part his article deals with the phenomena in the reduction-divisions
of Canna aureo-vittata gigas which was obtained by crossing Canna
aureo-vittata deep yellow x pale yellow. This new type is a real gigas
form, as well from the genetical point of view as from the cytological
one, for it has, either a) the double number of chromosomes or h) nine

very big (tetravalent) chromosomes (in the metaphase I) or c) in
other pollen mother cells uni-, bi- and tetravalents mixed. The C.
glaucaxindica Fj hybrid also often shows univalents in diakinesis.
In a few cases Honing found chains of chromosomes with differ-
ent forms. He leaves undecided the method of pairing of the chromo-
somes in early prophase, either para- or telosyndesis.

Kracauer (1930, „Die haploidgeneration von Canna indicaquot;)
describes the development of pollen and embryosac. With regard to
the meiosis of the pollen mother cells he says: „Die heterotypische
Teilung des Kernes geht in ihren Phasen völlig regelmäszig von
stattenquot;. This does not correspond with the words of Tokugawa and
Kuwada, quoted above, but Kracauer means probably the course
of the development in general and not details of the prophase. He
evidently investigates a regular strain, with n=9 chromosomes:
„Während aller Teilungs-phasen sind die Chromosomen in der
Haploidzahl (neun) mit Sicherheit zu zahlenquot;. It is a pity that
Kracauer does not give details about the prophase, for according to
his own words the development of the embryosac mother cells pro-
ceeds slowly from the one stage to the other.

The figures of his paper are too small to give a good idea of the
structure of the nucleus contents. Kracauer is of opinion that in the
embryosac mother cells a continuous spireme has developed, which
splits lengthwise in the synapsis (synezisis). He observes that the
gemini in diakinesis are not regularly scattered through the nuclear
cavity and that they do not form chains in this stage.

Observations are chiefly concerned with the reduction divisions
of the pollen mother cells; sometimes somatic tissue of the anthers
afforded good opportunities for counting chromosomes.

The inflorescenses of Canna are panicles (composed of cincinni)
which develop terminal and in the axils of the higher leaves of the
floriferous stalks; as a rule two flowers are close together, sometimes
with a small rudimentary third flower. If a shoot is already too far
developed, flowers containing the desired stages of the pollen mother
cells are often found in the secondary one which is still inside the
leaf.

The inflorescences were cut off and immediately put into water;
the stamens were prepared and fixed in the laboratory as soon as
possible. The Canna flower (fig. 2) has only one half stamen fertile
(fig. 3). In order to avoid the collection of unsuitable material, a
preliminary inspection of the contents of a little piece of each anther

was made in aceto-carmine (in accordance with the indications of
Belling, 1921, 1926, 1928, 1930, Heitz, 1926 and de Meijere
1928, 1930).

Later on in order to be able to state exactly the transition of the
different stages in the preparations, young complete inflorescences
were fixed in the greenhouse. I kept the anthers under the fixative
at the bottom of the tube by means of a dot of cottonwool; this turned
out to be not only a simple method but also a good one. The super-
fluous parts of the flowers were removed later on as much as possible
in the course of the dehydration. These complete inflorescences were
put into paraffin.

The following fixatives were chosen: Carnoy, Flemming (strong
mixture), Bouin-Allen and modified Nawaschin (of. Leliveld
1928).

Bad results were obtained with strong Flemming, because of con-
siderable shrinkage and distortion.

Carnoy's fluid is suitable for all stages; it may be an excellent
one for the prophase.

Bouin-Allen's fluid is generally a bad fixative fortheleptophase,
good for diakinesis, meta- and anaphase.

Nawaschin's fluid (modified) is fairly good for all stages, but like
Bouin-Allen's fluid, it gives a coarse precipitation of the nucleus-
contents in the early prophase. The chromosomes in meta- and
anaphase are more swollen and lumped together, which reduces the
possibility for countings and further study.

The imbedding in paraffin wax was first done in the usual manner.
Later on I proceeded to a modified method which in general may be
compared with those of Baker (1933).

As a rule the blocks were cut 18[i, thick ; very young inflorescences
sometimes 15jj. and to increase the possibility of getting sound nuclei
sections of 20[x were also made. The greater part of the material was
stained with Heidenhain's iron-haematoxylin, which gives an ex-
ceptionally clear staining of the chromosomes and their chromomeres.
The staining with Newton's gentian-violet gave satisfaction for the
chromosomes in meta- and anaphase, but not for the earlier prophase
stages. Comparing the stainings (on slides with sections from the
same anthers and with the same stages) the haematoxylin is
certainly preferable.

All drawings were made with the aid of a drawing-apparatus (at
the level of the work-table); objectives Zeiss H.I. 90x and 120x,
Fluorit 100 X with the compensating oculars : 10 x, 15 x and 20 x .

vi. description of the meiosis of the pollen mother cells
A. Canna humilis Bouché

The following description of the development of the pollen mother
cells starts at the moment when they are clearly to be distinghuished
from their surroundings, i.e. when the differentiation of tapetum and
archesporium has been accomplished. When this stage has been
reached, tapetal cells and pollen mother cells are conspicuous by
retaining stain in a higher degree than the cells of the surrounding
layers. The transverse section of an anther shows two locules with an
epidermis of one layer of square cells with large nuclei ; beneath that
about seven layers of parenchyma-cells follows, of which the outermost
rows show a more rectangular shape. (See for a metaphase plate of

these cells fig. 4). They become gradually smaller in radial diameter
in the direction of the inside. The innermost tapetal cells already very
soon contain more nuclei. Here and there in this
stage the (mitotic) division of the tapetal cells can
be seen. Sometimes very large tapetal cells are to be
found with a great number of chromosomes, proba-
bly originated by imperfect division. The develop-
ment of the tapetum and the degeneration of the
tapetal cells are no safe indications for the stage ofnbsp;^««»a

The archesporium (fig. 3) generally consists of two rows of large
multangular cells, the pollen mother cells. As a rule one is as far
developed as another in the same loculus; if any difference is to
be seen in development at all, then generally the pollen mother
cells in the upper part are in advance of the others. With fixation in
Bouin-Allen's fluid we get the following impression: The pollen
mother cells have very large nuclei in which a great dark-coloured
nucleolus is immediately prominent; usually the latter lies some way
off the nucleuswall and somewhat excentrically. The greater part of
the nucleus-hole is clear; it is filled with a fine reticulum of faintly
stained chromatin-threads. The more precise structure of these
threads which are not quite straight could not be accurately examin-
ed. In the nuclear cavity a number of small round bodies are to be
seen, which have assumed the same colour as the nucleolus; regularity
in the appearance of these small bodies, as regards size or number
could not be observed (fig. 5).

The granular substance of the reticulum becomes gradually more
clearly visible, the granules become more darkly coloured and larger;
the threads themselves are very thin, sometimes hardly visible. Later
on they are directed more meridional and often run parallel, but it is
not possible to see a pairing in this phenomenon. Free ends in sound
nuclei are not to be seen (fig. 6).

This stage is followed by a contraction of the reticulum with the
coherent chromatic material in the vicinity of the nucleolus: obvi-
ously this does not happen suddenly, but gradually, for transitional
stages are to be found in the preparations. (The. typical synaptic knot
is an artefact, undoubtedly, owing to bad fixation).

Further advanced stages are to be found in the same preparations:
in the nuclear cavity a clear thread is to be seen, lying in large loops.
It is impossible to say whether we must speak of a continuous spireme

Fig. 5. C. humilis, Nawaschin,
early-zygophase, 1800 X.

or of separate loops; free ends are not to be seen. At a later stage
the threads contract ; their smooth sides disappear and they become
thicker and get a more granular surface. Sometimes the threads
show signs of doubleness in that their margins are darker and the
centre lighter.

As it appears from the following explanation the stage just
mentioned is zygotene-pachytene. On comparing preparations, which
were made with different fixation-fluids, it appeared that the fixation
with Bouin-Allen's or Nawaschin's fluid gave an incorrect idea of the
structure of the nuclear contents in the early prophase stages. Both
fixation-fluids mentioned, as a rule, give too coarse a precipitation, so
that the finer structure, namely of leptotene-threads and of the
chromomeres is totally absent. On account of this the actual pairing
of the leptotene-threads and the exact development of the zygotene-
stage becomes invisible. It is not superfluous to point out once more
that the aspect of the chromosomes may be to a great extent de-
pendent on the fixative used ; different fixatives give rise to different
figures in the same stage.

researchers have altogether different opinions about the development
in the same material, especially with regard to problems, such as
those of the para- and telosyndesis. Now in those cases, where good
results were obtained with Canna humilis when fixed in Carnoy's
fixation-fluid, the leptotene-stage presented no difficulties in this
respect, so that the development of the bivalents could easily be
traced. Moreover they were of importance for the explanation of the
figures of the chromosomes, obtained by using the other fixatives
mentioned. It is conspicuous that the nucleolus is not so dark

coloured as after using Bouin or Nawaschin; it seems to be a globular
body of alveolar construction. Similar large chromatic bodies are not
present (with the exception of one small round body in the immediate
neighbourhood of the nucleolus).

By conscientiously studying this stage it can be concluded that the
very fine threads associate in pairs; in favourable circumstances the
chromomeres of two threads are seen to be lying in corresponding
places (homologous threads with homologous chromomeres, fig. 7).
The pairing threads contract, lengthwise, on account of which the
chromomeres approach each other more and more, for the developing

thicker thread is smooth and dark-coloured. Usually leptotene-
threads and zygotene-threads are found to be lying together in a

nucleus; clearly distinguishable by their appearance, namely a) thin
threads with fine chromomeres at unequal and fairly large distances,
and b) thick, dark-coloured threads with chromomeres united into a

continuous row (fig. 8). Thiis gives the so-called amphitene of
Janssens. This stage passes into the ultimate zygotene; it produces
the aspect of the „open spiremequot;, large loops he throughout the
nuclear cavity (fig. 9 and fig. 10). In early zygotene the threads show
light-coloured or nearly colourless spots (fig. 11); it is not impossible
that these „gapsquot; become the places for the spindle-fibre-attach-
ments (attachment-constrictions), but their further development
was untraceable here. Obviously the threads of this „open spiremequot;
contract. (Whether this is a „continuousquot; spireme or not, cannot be
definitely stated; free ends are not to be seen.)

They grow thicker and often come to lie closer together; their
sides become corrugated, i.e. more clearly granular. In the middle of
these threads a lengthwise split becomes visible; the double structure
of the threads is apparent. What has happened to the chromomeres
or the chromatids, cannot be perceived. It cannot be stated with
certainty that in this stage four chromatids constitute the thread.
The pachytene-loops expand and fairly suddenly the nuclear cavity
is filled up with the diplotene threads (fig. 13). Now the chromomeres
are visible again, situated on twisted and crossed paired threads.
At this stage it is not possible to study the whole complement, for the
chromosomes are so long, that they cannot be individually traced; it
stands to reason that the connections of the chromatids, the chias-
mata, are not visible either. If the crossings and twists of the threads
are to be considered as the points where the chiasmata are formed, it
is probable that these are formed at random. The precise number of
chiasmata is not to be ascertained (but there may be five or six to
each bivalent). This stage changes into diakinesis in a remarkable
manner for the contraction of the bivalents does not take place along
their whole length simultaneously. The contraction of the diakinesis-
bivalents is gradual; when local condensation has already materially
advanced, the remainder is still in the diplotene stage (fig. 12a, 12amp;
and 14 a—i). At the same time both the diplotene and the diakinesis-
stage are observable, namely, the various bivalents of the same
complement do not contract synchronously which is neither the case
with the development of certain portions of each bivalent. This
remarkable phenomenon might be called heterochronous conden-
sation of the bivalents (fig. 15«, 156). Although it cannot be as-
certained accurately how many chiasmata were originally formed, it is

clear, that their number gradually diminishes (terminahzation) ;
see fig. 16, 17, 18 and 19. Finally the chiasmata are all terminal.
At the end of diakinesis the condition of the bivalents is nearly
uniform, for most of them are ring-bivalents, except one or two,
which are rod-bivalents (fig. 18). The „end-to-end associationquot; of
these bivalents has arisen owing to the „terminahzationquot; (probably)
of the chiasmata. When pursuing the development of the bivalents
in diplotene, the impression is created that the chromatids in each

pair are spirally twisted (fig. 12,14-17). They contract, the spirals be-
come shorter and thicker; finally both homologous parts of the
bivalents give the impression of a double thread. Ultimately the
bivalents are totally condensed into nine double bodies (fig. 20).
Meanwhile the nucleolus decreases in size, whilst at the same time a
smaller body of a similar appearance becomes visible in the immedi-
ate neighbourhood of it (probably there is some connection between
them). When the nuclear cavity disappears, the spindle figure be-
comes visible (fig. 21). Possibly this does not take place instan-
taneously, though transitional stages are rarely to be seen. A multi-

polar spindle could not be observed. After having first been scattered
in the spindle (fig. 22), the bivalents are drawn up in the centre.
Sometimes they lie in a cluster (fig. 25) before they assume their
position in the metaphase-plate (fig. 23). The metaphase I is very
suitable for counts; in several preparations this stage is frequent and
moreover the bivalents are clearly distinguishable in polar views.

They are equal as to form and size. In anaphase I the individuals of
the bivalents separate; this is a quick process (fig. 26). The transition
from metaphase I to anaphase I is in case of Canna humilis not
very suitable for observation of the remaining chiasmata.

From the diakinesis the conclusion may be drawn that a complete
terminalization is brought about (or, at any rate, is probably brought
about), but in early anaphase I the chromosomes are too small and
too compact, to confirm the above statement. Separation of the
chromosomes in anaphase I is not always synchronous; one or two
may be somewhat in advance of the others. They may possibly have
arisen from the rod -pairs. This too cannot be proved with absolute
certainty, as the closely associated bivalents in metaphase I are not
distinguishable (fig. 23). Neither is there any striking difference
between the chromosomes of an anaphase-complement (fig. 26«, b).
A great number of corresponding anaphase-plates are found in the
preparations; the separation in 9 9 is very regular (fig. 26, c—/). In
clear cases the anaphase chromosomes are seen to be double; their
aspect changes from early to late anaphase (fig. 27). At first four
free extremities are visible, connected at one point at least; later on the
ends are drawn together and a lengthwise split is conspicuous. This

is still the case in telophase I; the chromatids which are connected
at the ends and at a point towards the middle are clearly distinguish-

able ; this gives the impression of an eight-shaped figure, which is
slightly spiral-formed (fig. 28).

The spindle has in the meantime become more globular and has
expanded in the width (phragmoplast) ; in the middle arises a new
cell-wall, separating the two cells of the dyad. This takes place as
usually in monocotyledonous plants (fig. 29—33). Occasionally nine
chromosomes may still be easily counted in the nuclei of the dyads
(fig. 30) ; they are often constricted in the middle. As a rule vacuol-
ization soon appears, so that the individual chromosomes are no longer
distinguishable. Little can be said about the duration of the inter-
phase ; sometimes this stage is exclusively found in both the loculi of
an anther, at other times dyads and tetrads are found together, with
all the phases of the second meiotic division. In metaphase II, the

long and narrow spindles (fig. 29) often lie parallel or nearly parallel
with their long axes, but they may also cross each other. The factors,
causing these different positions, are not to be inferred from the
preparations.

seen that the chromosomes separate regularly in two sets of 9 each
(fig. 32). Very often, however, the fairly small chromosomes are
grouped together, offering difficulties for observation. The tetrads
arise in the usual manner (fig. 29, 33).

There are abnormalities (e.g. fig. 34, an abn. spindle-form), but
in such a small percentage that they are of no significance for our
considerations. The pollen is formed abundantly and only a very
few pollen-grains are sterile.

The zygophase and pachyphase develop as described for Canna
humilis. If differentiation is correct, in late pachyphase details are
observable in the threads, which give the impression, that these
really consist of four strings (fig. 35). This stage passes into diplotene
(fig. 37); with this species too it is not possible to state with certainty
the development of the chiasmata and their exact number, firstly

because the chromonemata are extremely hard to trace, secondly on
account of the fact that the contraction in the bivalents locally is
much further advanced than elsewhere (fig. 36 and 38). The whole
process proceeds gradually, but not quite simultaneously for all the
bivalents in the same nucleus. Finally 9 bivalents are developed,
which contract more and more. The paired chromatids of a bivalent
are obviously spirally coiled round each other. It is clear that only
terminal chiasmata remain; a single rod-bivalent is developed.

The transition from prophase I to metaphase I etc., and the second
meiotic division can be a quick process, for all these stages were
found in a single anther. The process is, as a rule, very regular; in
most of the cases observed, the chromosomes separate in equal
number (9 9) to either pole (fig. 39). The dyads look very normal,
as is the case with the tetrads. Sometimes there are deviations: a) di-
visions with lagging chromosomes and b) divisions, which give rise to
unequal numbers. What may be the outcome of these abnormal di-
visions, remains to be seen. The abundantly formed pollen looks very
normal; conspicuous abnormalities are certainly absent.

As already mentioned (in the introduction) the variety examined
is presumably of hybrid origin; it has a regular development as to the
meiosis of the pollen mother cells. The stages of the prophase are
essentially similar to those of Canna humilis and the other form de-
scribed. The leptotene-threads associate in pairs synaptically, (fig.
40); the transition from one stage to the other proceeds gradually,
especially the condensation of the bivalents. With Canna discolor
meiotic stages may be found where some of the chromosomes form
chains, but it is a rare phenomenon; moreover it is very well possible
that in most of the cases such a figure will be a „pseudo-chainquot; (only
optically a chain) with the bivalents lying close together in a straight
or in a curved line. In metaphase I connected chromosomes are not to
be seen: nine bivalents are distinctly developed (fig. 41—42). The
separation of the chromosomes in anaphase I proceeds regularly; the
splitting of the anaphase-chromosomes is evident and their form
changes from early to late anaphase,, as already described for the
other species (fig. 43).

Fig. 40. C. discolor, Zygophase, 1800 x .
Fig. 41. C. discolor, Anaphase I, 3600 X .
Fig. 42, C. discolor, Metaphase-plate I, 3600 x .
Fig. 43. C. discolor. Anaphase I, late, 1800 x .

(The meiosis in the varieties „Javaquot;, „Boliviaquot;, „Montevideoquot;
and „Pure yellowquot;.)

Somatic division : Frequently, somatic metaphase plates in the
parenchyma cells of the wall of the anthers could be observed;
countings resulted in 18 chromosomes (fig. 44). Their finer morpholog-
ical structure could not be studied, for the fixation-fluids used,
were not suitable for that purpose. Notwithstanding, in the same
anther e.g. the nuclei of some tapetal cells can show very fine fixed
chromosomes with clearly visible constrictions and trabants. Owing

to their minuteness the chromosomes in these nuclei are not suitable
for further examination (see fig. 45«,amp;, showing a nucleus of a
somatic cell and a pollen mother cell at the same magnification).

No morphological differences were to be found in the somatic
chromosome-complements of Canna glauca „Javaquot;, „Boliviaquot;, „Mon-
tevideoquot; and „Pure yellowquot;, but it is probable that with a closer
examination, after using other fixatives, they may be detected (as
regards size, shape, constrictions and trabants).

Meiotic division. Prophase I: The prophase stages in these types
are very similar to those described already for Canna humilis. The
following descriptions apply to

If penetration was good and the fixation was satisfactory, the
chromomere-structure of the leptotene-threads is evident. The
minute tortuous threads themselves are hardly visible; the dark

stained granules lie at unequal distances in the slightly stained
threads. Both granules and threads are single (fig. 46).

These leptotene-threads associate in pairs side by side, which is to
be concluded chiefly from the appearance of the^so-called amphitene-

stage; it is extremely difficult to determine when and where these
delicate threads first come into contact with each other, but it is
certainly a gradual process. In many cases the chromomeres of one
thread are situated directly opposite the chromomeres on the other
thread, so that this is a clear case of parasynaptic pairing (fig. 47).
There is a normal zygotene and pachytene stage following this as-
sociation (fig. 48-50); then the noticeably thicker thread can be
seen in some places to be double, but four chromatids are not dis-
tinguishable. How the precise transition from pachytene (fig. 51)
to diplotene (fig. 52) takes place is obscure. Apparently it is a rather
rapid process, for suddenly the nucleus is filled up for the greater part
with the diplotene threads. Sometimes their finer construction is to
be examined; obviously they are tetrads, consisting of four crossed
and twisted chromatids (fig. 53). At the same time the partly con-
densed bivalents are already visible, for the condensation does not
take place simultaneously. Each of the two chromatid-pairs of a
bivalent has an already condensed portion, which lie in homologous
places; the other parts, still in the diplotene-stage, lie in different
directions in the nuclear cavity. Only in a few cases these threads
can be traced over a greater distance (fig. 53-54). By the further
contraction of the bivalents (which in this species is a heterochronous
phenomenon too) the real diakinesis develops; nine bivalents occur
whose double nature is still quite distinct (fig. 55-56). Generally
most of them are ring-bivalents; sometimes one or two rod-bivalents
are developed (as was the case in Canna humilis).

In a few nuclei associations of two bivalents have been found
(seefig. 57, 58, 59).

It cannot be decided how they have been formed; for the con-
nections between the pairs are not quite clear; it may be in the one case
(fig. 58) that two bivalents, which were totally terminalized, have
come to lie parallel. If this is so, it is not a very interesting phenom-
enon. But such is by no means the case in the association shown by
fig. 57 and 59, for here are two bivalents evidently connected by a
fine, probably, double thread. It may be considered as a „secondary
associationquot; (but this is merely a word here and not an explanation).
At the end of diakinesis the nucleolus disappears; the nuclear cavity
decreases. The now highly contracted bivalents lie scattered through
the nucleus (fig. 60) sometimes in one half of the space and at a short

distance from the nucleus membrane; certainly they are not of the
same form and size, but it has not been possible to distinguish the
individuals of the complement. In this stage and the following, when
the bivalents are scattered through the spindle meanwhile developing
(fig. 61-62), they may be easily counted. With Canna glauca
„Javaquot; the chromosomes in the bivalents are closely associated,
although as a rule the double nature of the latter is to be concluded
from their curved outline.

Sometimes there is a contraction following diakinesis : the chromo-
somes clump together in the centre of the pollen mother cell.

So far as may be judged from the observations, the spindle is
a product of the nucleus; I never found in the immediate sur-
roundings of the nucleus any indication of a multipolar spindle. But in
several cases fine threads are to be seen in the nuclear-hole, which
becomes more bipolar. These threads bear in some places small
stained particles (with the same colour as the chromatic material). It
is impossible to say, whether they are still parts of the chromosomes,
or perhaps „proto-fibresquot; of the spindle.

Metaphase I : The metaphase-plate is regular, but it is rather rare
in the preparations, owing to the probably short duration of this
stage. From side and polar views it may be inferred that the bivalents
have terminal chiasmata, one or two, for they are ring- or rod-
shaped. The primary constrictions are seldom quite clear.

Anaphase I: Although the separation of the chromosomes in
early anaphase is not always simultaneous, this stage cannot be
said to be irregular. Some chromosomes, possibly the rod-bivalents,
may move apart earlier or more quickly (fig. 63). In this stage, counts
offer no difficulty, neither in side-view nor in polar view ; especially
the latter give, in many cases, the impression of similarity in the
grouping of the separating daughter-chromosomes. Fig. 64«, where
the chromosomes are clearly seen passing towards the poles, gives a
good notion of the anaphase I : a) the chromosomes separate regularly
in 9 9; h) the double nature of them is evident (from early to mid-
anaphase) ; c) not all the chromosomes of a set are in the same stage
of development. In late anaphase the chromosomes are evidently
split, but their appearance is somewhat different from that in early
anaphase.

mmfiim.

liii

i ■■ ! :
^ 7 /

î-^imtf^Wf' 63

Fig. 61. C. glauca „Javaquot;, Chromosomes scattered through the spindle
3600 X.

Fig. 62. C. glauca „Javaquot;, a-b-c: chromosomes scattered through the spindle,
2700 X.

Fig. 63. C. glauca „Javaquot;, Metaphase I, 3600 x .
Fig. 64. C. glauca „Javaquot;, a-h, Anaphase I, 2000 x .

come in contact with one another, sometimes forming a spiral-curved
chain (fig. 65). Since they are rather large, counts again are easy to
make, but they soon lose their sharp outlines by alveolization and
the individuals are no longer distinguishable. In a transitional stage

from anaphase to telophase they are evidently double, for a split at
the length can be seen, while probably the ends of the chromosomes
are connected. The spindle between the daughter-nuclei extends

ai
Oi

bC
cS
IH

laterally as a phragmoplast. The first meiotic division ends with
the formation of a new cell-wall, as is the rule in monocotyledonous
plants (fig. 66, 67, 68b).

Second meiotic division: The cell-wall, separating the two
daughter-cells is sometimes hardly visible; superficial examination
may give the impression that no membrane at all exists.

Metaphase and Anaphase II: The spindles can he parallel with
their axes, but there are also cases where they he cross-wise (fig. 67a
and c). The causes governing such differences in position are obscure.
These spindles are to be distinguished from the spindles in the first
meiotic division by their shape, viz. long and narrow. A regular
distribution of 9 chromosomes to each pole has been observed (fig.
69a). Irregular separation of chromosomes must be rare. In metaphase
II the chromosomes on side view are ring-shaped (fig. 67b); on top-
view the anaphase chromosomes have a remarkable appearance, for
it gives the impression that there are too many (when namely both '
arms of each chromosome are clearly visible and the connection
between them is vague; fig. 69).

The anaphase proceeds very regularly and changes into telophase
II in which the chromosomes clump together and form a dense mass
at the poles of the spindles. As a rule the new membrane develops in a
regular manner and the tetrads too. If there are any irregularities
in the formation of the tetrads they are rare and probably due to
marring external circumstances and not to internal factors.

Concerning the development of the pollen mother cells of Canna
glauca „Boliviaquot; the following facts may be mentioned: The diploid
number of chromosomes = 18 (metaphase plates in somatic cells of
the anthers could easily be counted, fig. 70).

After the resting-stage the nucleus passes into the leptotene-stage.
The chromomeres are very clearly visible; the synaptic pairing of the
leptotene-threads is evident. Zygophase-pachyphase develops as
already described for C. glauca „Javaquot;. The diplotene-stage passes in
following manner into the diakinesis: The bivalents condense
gradually; the contraction begins in a certain portion of the bivalents
(but in homologous parts) and this proceeds succedaneously with
the still projecting diplotene threads. These spirally wounded
threads contract more and more (fig. 71,72 and 79).

traction, which in this form is a heterochronous phenomenon too (fig.
72). The parts of the bivalents already condensed often He together
in another part of the nucleus than the attached non-condensed

..^^iiaiiiiii^

f
c»»

Fig. 76. C. glauca „Boliviaquot;, Chromosomes, scattered through the spindle,
a-b-c, 3600 X .

Fig. 77. C. glauca „Boliviaquot;, Anaphase I, a. 1800 X , b. 3600 x , c. 3600 x .

threads ; the connection of bivalents and threads is clearly visible in
„Boliviaquot;.

At the end of diakinesis nine gemini have been developed ; three of
them may be found in several cases near the nucleolus, while the

others he closer to the nucleus-wall (fig. 74). The metaphase-chromo-
somes have as a rule two terminal connections; one of these con-
nections may become detached first, the other being more tenacious.

When the chromosomes are scattered through the spindle, their
double nature is evident although the chromosomes of each bivalent
lie close together (fig. 75). The gemini may show a „secondary associ-

ationquot; (fig. 76a). The spindle may be found somewhat excentrically.
Obviously the transition from diakinesis to anaphase proceeds very
rapidly.

The separation of the chromosomes in anaphase is regular: 9
chromosomes pass to each pole, but this does not take place simul-
taneously (fig. 77). In mid and late anaphase the chromosomes are
clearly double; in late anaphase they are somewhat spiral-formed.
For the rest see: C. glauca „Javaquot; and the illustrations, belonging
to C. glauca „Boliviaquot; (fig. 78a and b).

This variety requires no detailed description for the same results
were obtained from the investigations as with the forms already
described. The pairing is evidently parasynaptic (fig. 80).

The diploid number is 18. (Somatic cells of the anthers, fig. 81).
Clear differences with the other forms examined, as regards the meiot-
ic phenomena, are not to be seen.

Undoubtedly, in the somatic tissue of the anthers the metaphase
plate clearly shows 18 chromosomes (fig. 82). It is very probable
that morphological differences occur, but owing to the unsuitable
fixing fluid they were not so clearly developed that details could be
analyzed. The preparations available did not show in the early
prophase the fine chromomere-structure of the leptotene-threads,
but we may venture to suppose that the zygotene-pairing takes place
in the same manner as it has already been described for Canna humilis
and the other forms. (For the appearance of zygotene-pachytene is
essentially the same as that of the other forms described). Apparently
these stages pass gradually from one to the other, but not simul-
taneously for all the chromosomes of a pollen mother cell.

The ultimate zygotene threads, lying in large loops in the nuclear
cavity are undoubtedly double: often they lump together in the
proximity of the nucleolus. The pachytene and the diplotene develop
in the manner mentioned. The individual chromosomes are not to be
traced, so thatja detailed study of this stage, namely accurate count-

Fig. 93. C. glauca x discolor F„ Seven bivalents and four univalents, 3600 X .

ing of chiasmata is impossible (as in so many other cases). The
chromosomes become distinguishable at the beginning of diakinesis,
while in mid diplotene the condensation of some bivalents is already

Fig. 94. C. glauca x discolor F^, Anaphase I, similar grouping of the chromo-
somes, ± 3600 X .

Fig. 95. C. glauca x discolor Fj, Late anaphase I, 2400 X .
Fig. 96. C. glauca X discolor Fj, Mid anaphase I, 2400 X .
Fig. 97. C. glauca x discolor F„ Multivalent connections.

clearly visible (fig. 83); in late diplotene the greater part of the
threads again are seen to be double and paired in most of the nuclei

Ï/
\ .V?

Fig. 98-99. C. glauca x discolor Spindle-development in late diakinesis.
Fig. 101. C. glauca x discolor Abnormal spindle.

(fig. 84, 85). But generally the exact analysis of the nuclei is
difficult, owing to the fact that the gemini as a rule are found in the
same part of the nucleus, so that they often cover each other, wholly
or partially. In the course of diakinesis the association of the chromo-

somes gives the impression that nine pairs will be formed (fig. 91 ) ; it
may be that some are incompletely paired for they approach each
other without uniting (fig. 93, 88). Univalents lying wide apart are
to be seen in a few cases only. In an advanced stage of diakinesis
pollen mother cells sometimes still contain a great number of prob-
ably unpaired threads ; the chromomeres are larger than those of the
leptotene threads. Possibly the former threads may develop into
univalents. In metaphase I pairing is almost complete, although in

Fig. 100. C. glauca x discolor, Metaphase and anaphase spindles of the
second meiotic division, 1800 x.

Typical metaphase plates with nine apparent bivalents are seldom
found; very soon some of them are separated (fig. 90, 89).

In diakinesis surely multivalent connections may be developed,
but this is a rare occurrence (fig. 86, 87, 97 a-e). Typical chains are
surely very rare.

When the bivalents are arranged end-to-end but without a real
connection, which is quite well possible (a portion in fig. 93), they
do only form a „pseudochainquot;, so that this phenomenon might be
called „pseudo-catenationquot;. Two bivalents in late diakinesis or in
metaphase I are often found strikingly large; probably they are
glauca-bivalents. Not seldom seven bivalents and four univalents
are found (fig. 86, 87, 93).

In the material available comparatively few clear early anaphases
were to be found; in those which were found, the chromosomes
regularly separated 9 9 (fig. 92, 94). In order to make sure, many
telophases were examined (fig. 95) ; all these gave the same results.

If one has a clear notion of the structure of the chromosomes in
late anaphase and early telophase, they may easily be counted.
Their structure is equal to that of the chromosomes of the different
Canna types described in the same stages (fig. 96, 95).

The further development follows the usual course and needs no
more elucidation (fig. 100; metaphase and anaphase II).

Terminology. Present-day terminology for the prophase-stages
is not entirely suitable for the facts described here. Properly speaking,
the terms leptotene, zygotene, etc. cannot be used (in many cases)
in their original significance, seeing that these stages are often not
distinctly separated, but gradually change from one into the other.
The difficulty of denominating the transitional stages has already
been felt by many writers which has given rise to expressions
such as amphitene, mid-zygotene, pro-metaphase, etc.

With Canna, there is no definite zygotene stage for the whole
nucleus; often paired and single threads are to be seen in the same
nucleus, a stage, equivalent to amphitene of Janssens and to mid-
zygotene of other researchers. The words of Huskins and Smith
(1935): „And pachytene likewise is not a definite stage for the whole
nucleusquot; are applicable for Canna too. The same holds good for
diplotene and diakinesis.

It would be recommendable to choose new terms, more expressive
of the dynamic character of the phenomena, but present-day termin-
ology has become too popular in cytology to admit of other terms.

Fixation. Although some researchers object to the use of Car-
noy's fluid (Darlington, 1932, p. 491) it was successfully used in
this investigation and late literature creates the impression that a
great many cytologists again resort to the use of this fixative.

On comparing the results of investigations notice should be taken
of the fixation-fluids used for the treatment, as it has appeared from
this investigation, too, that the results may be very different, owing to
the fixatives used (p. 18). If the above is not taken into account it
may lead to misconceptions, especially of the finer structures. It is
possible that in some cases wrong conclusions were drawn, e.g. with
regard to prochromosomes and telosyndetic pairing, owing to un-

suitable fixatives. It is worth mentioning the words of Gates and
GooDVi'iN (1931): „The papers of Latter (1926) and of Maeda (1930)
on Lathyrus, also lead to the conclusion that according to the
method of fixation one may obtain an essentially telosynaptic or
parasynaptic story from preparations, the fusion of delicate parallel
threads in certain fixatives giving an appearance of telosynapsisquot;.

Prochromosomes. In the pollen mother cells of the species
examined no prochromosomes could be detected, (neither in the
embryo-sac mother cells observed) although, on the contrary,
Kracauer (1930) does state the existence in the embryo-sac mother
cells of Canna indica, but he does not say so of the pollen mother
cells.

Method of pairing. Definite evidence of parasynapsis in Canna
has been found. The pairing of the homologous chromosomes in
Canna does not take place simultaneously along their whole length,
a phenomenon which is also observed with other plants: some threads
being partly paired and partly unpaired, the latter conspicuously
thinner than the paired ones.

It is open to doubt, whether contact originates accidentally in any
place or whether it is a regular process. Counting or estimating the
number of chromomeres in a nucleus, which has appeared to be
possible for species of Lilium and Trillium (Belling, 1928, 1931;
Huskins and Smith, 1935), is impossible with Canna.

That a continuous spireme develops is not excluded, yet highly
improbable. When the chromosomes in early prophase are long,
as is the case with Canna, free ends are difficult to observe. In
favourable cases the loops in pachytene gave the impression of
consisting of four strings; it is, therefore very probable that in this
stage four distinct chromatids are developed.

Condensation. In what manner the pachytene changes into the
diplotene-stage is not quite clear. In particular, it has not been possi-
ble to discover with certainty the development of the connections
between chromatids and chromatid-pairs, i.e. the formation of the
chiasmata. Indicating the places and the number of the chiasmata
and the change in their position, is very difficult or hardly possible,
as the bivalents do not contract simultaneously along their whole
length. Locally this process is more in advance; probably this more
condensed part does not lie in the middle, but (at any rate for some

of tfie bivalents) more towards one of the ends, the distal parts are
then still in the diplotene stage. There is a relation between the
quantity of diplotene threads and the contracted parts of the bi-
valents. This very gradual decrease of the mass of diplotene threads
and the gradual increase of the condensing portions of the bivalents
is probably not so conspicuous in other species.

(Surely this phenomenon cannot be the consequence of bad fix-
ation, as it appears after different treatments and with different
species and varieties, at different times collected).

Only occasionally similar observations are met with in literature,
but these phenomena are however not quite the same as those de-
scribed for Canna.

Catcheside (1931) describes a comparable phenomenon for
Oenothera: „the chromosomes have the major portion of their
substance located between two chiasmata, suggesting that ( 1 ) chiasma
formation is most frequent towards the ends of the bivalents, and
that (2) there is a movement of chiasmata away from this portion
towards the much more slender distal portionsquot;. „They are very
variable in thickness and outhne, being thick at their centres and
thinner distally, the surface in the thick portion is more or less ir-
regular, while in the thinnest distal portions an appearance suggestive
of a row of variously sized chromomeres may be distinguished in
some of the less condensed bivalentsquot;.

Latter (1932) in a description of the meiotic divisions in the pollen
mother cells of Malva sylvestris mentions, that an extreme con-
traction of the bivalents occurs immediately prior to diakinesis, and
that the appearance of the nuclei in this stage suggests that a portion
of the thread mass is not included in the chromosomes. In Malva
sylvestris these remaining threads form a faintly staining mass of
somewhat reticulate structure around the nucleolus, while the con-
tracted bivalents take up a peripheral position. In all diakinetic
nuclei a quantity of this faintly-stained substance as well as the
bivalent chromosomes has been observed by Latter. But compara-
tively few stages of diakinesis were found in the material examined.
Latter supposes that this remaining portion probably corresponds
to the linin-threads, described in Lavatera by Byxbee, which, ac-
cording to her account, contribute to spindle formation. In Malva
sylvestris these faintly-stained indistinct strands of substance are

present in the nucleus in addition to the bivalent chromosomes still
in late diakinesis; according to Latter they w^ere not utilized at
chromosome-formation. With Malva sylvestris there is evidently no
clear connection between the threads and the bivalents in diakinesis,
the structure of the threads cannot be inferred with certainty from
the illustrations. With the species of the genus Canna examined, the
threads mentioned are gradually taken up into the chromosomes and
therefore the phenomenon is comparable with that in Oenothera, as
described by Catcheside.

Hugh Davie (1934) studying Lavatera, found also, that in
diakinesis the appearance of the nuclei suggests that not all the
chromatic material has been used up in the formation of the bivalent
chromosomes.

Spindle. Byxbee supposes, the threads contribute to spindle for-
mation; Latter had insufficient material for an accurate study, but
yet he has a different notion of the mode of spindle formation. Byxbee
describes the cytoplasm of Lavatera as composed of two constituents,
a fibrous network and a granular substance about the nuclear wall,
which leads to the formation of the characteristic „perinuclear zonequot;.
The fibres grow into the nuclear cavity and mingle with the „linin-
threadsquot;. Surely this is not the case with Canna) the „threadsquot;
cannot be brought into relation with the formation of the spindle,
although in some cases it was clearly observable in late diakinesis
that the nuclear-cavity had become spindle-shaped and within
threads were visible, bearing small, stained particles (fig. 21, 98, 99).
A multipolar spindle or the inward growth of spindle fibres has never
been observed with the Ca««a-species examined. These observations
support the view that the spindle arises from the nucleus, but there
is no definite evidence, that it originates in a portion of the threads
with chromatic material.

Although rare, abnormal spindles have been found (fig. 34, fig.
101); in all cases the chromosomes he within the karyoplasm of the
spindle.

Sometimes, each chromosome in metaphase and anaphase seems
to lie in a separate part of the spindle (fig. 94a, b). Some investigators
consider this to be an artefact, but to others it is an indication that
the spindle is a compound structure consisting of elements, related to
the chromosomes. Akhough such cells and spindles look very normal.

this, however, does not mean that in hving cells the same structure
would be observable. It may be that the spindle-fibres must be
considered as artefacts (cf. Belar 1928; Bleier 1930; H. Davie
1934) and likewise it may be for the described anaphase-pictures.
(The illustration fig. 94« and h shows clearly the similarity of the
grouping of the chromosomes).

Associations. The metaphase of the first meiotic division suggests
that some of the chromosome-bivalents show secondary associations;
some figures in a paper of Honing, too, (Canna Crosses II), give this
impression. If there is actual secondary association, the whole
chromosome-complement is composed of 3 bivalents and 3x2 (more
related) bivalents and it must be possible, that some species are
polyploids. But there are no other indications in support of this
suggestion; species with a smaller number of chromosomes than
18 have not been found so far. Neither have particularities been
published about the pairing in triploid forms, which might elucidate
this phenomenon.

As a rule meiosis proceeds very regularly with clear formation of
nine bivalents, this is worth mentioning, as most of the cultivated
Cannas (and presumably also many so-called „wildquot; forms) are
species-hybrids.

Coupling of factors, found by Honing, cannot be ascribed to chro-
mosome-coupling, connections as rings or chains are of rare occur-
rence and it is quite well possible that only pseudo-chains are formed
(by secondary association).

It is probable that a comparative study of the development of the
diploid, triploid and gigas-forms may shed more light on the subject.

A. 1. With Canna humilis the chromosomes associate in pairs: it is
a clear case of parasyndesis.

2.nbsp;Though the chromomeres are clearly visible (after suitable
fixation) their number cannot be definitely fixed.

3.nbsp;It is probable that in the diplotene-stage interstitial chiasmata
are formed, which become completely terminalized.

4.nbsp;The condensation of the bivalents during diplotene-diakinesis
proceeds gradually and heterochronously.

7.nbsp;The haploid number is 9, which is clearly evident from diaki-
nesis, metaphase and anaphase.

In anaphase I and also in anaphase II the chromosomes sepa-
rate regularly in 9 9.

9.nbsp;A certain portion of the pollen mother cells perishes already
before division; the causes are not to be ascertained.

10.nbsp;Of the abundance of pollen formed, only a very small per-
centage is sterile.

B. The other forms require no further description here for
almost the same results were obtained from the investigations.

My best thanks are due to Prof. Dr. J. A. Honing for his interest
in my work and for valuable criticism in the preparation of this paper.

Canna. Proc. Nat. Ac. Ss. Washington, Bd. 7.
Belling, J., 1921. On counting chromosomes in pollen mother cells. Amer.
Natur. 55.

Belling, J., 1925. a. Production of triploid and tetraploid plants. J. Hered.
Vol. 16.

Belling, J., 1925. b. Chromosomes of Canna and of Hemerocallis. J. Hered.
Vol. 16.

chromosomes. Biol. Bull. 50.
Belling, J., 1928. A method for the study of chromosomes in pollen mother

cells. Univ. Calif. Pub. Bot. 14.
Belling, J., 1930. The use of the microscope. New York—London.
Belling, J., 1931. Chromomeres of Liliaceous plants. Univ. Calif. Pub. Bot.
16.

Bleier, H., 1930. Untersuchungen über das Verhalten der verschiedenen
Kernkomponenten bei der Reduktionsteilung von Bastarden. La Cel-
lule 40.

Byxbee, E. S., 1900. The development of the karyokinetic spindle in the
pollen mother cells of Lavatera. Proc. Calif. Acad. Sei. Ser. 3 (Bot.) 2.

Catcheside, D. G., 1931. Critical evidence of parasynapsis in Oenothera.
Proc. Roy. Soc., B. 109.

Costerus, J. C., 19160. A fresh investigation into the structure of the flower
of Canna. Ann. Jard. Bot. Buitenzorg. 2 Serie, Vol. XIV.

Costerus, J. C., 1916c. Die Übereinstimmung und der Unterschied in dem
Bau der Blumen von Canna und derjenigen der Marantaceen. Ann. Jard.
Botan. Buitenzorg. 2 Serie, Vol. XV.

Costerus, J. C., 1920. La structure de la fleur du Canna. Ree. trav. bot.
Neerl. XVII.

Davie, J. Hugh, 1934. Cytological studies in the Malvacae and certain related
families. Journ. of Genetics. Vol. XXVIII.

Gaiser, L. O., 1930. a. Chromosome numbers in Angiosperms II. Bibliogr.
Genet. 6.

Gates, R. R. and Goodwin, K. M., 1931. Meiosis in Oenothera purpurata
and Oe. blandina. Proc. Roy. Soc., B. 109.

Grégoire, v., 1912. Les phénomènes de la métaphase et de l'anaphase, etc.
Ann. Soc. Sei. Bruxelles. 36.

Heitz, E., 1929. Heterochromatin, Chromocentren, Chromomeren. Ber.d.d.
bot. Ges. 47.

Honing, J. A., 1915. Kreuzungsversuche mit Canna-varietäten. Reo. trav.
bot. Neerl. 12.

Honing, J. A., 1923. Canna Crosses I. Meded. Landb. Hoogeschool, Wage-
ningen. Vol. 26. Paper 2.

Honing, J. A,, 1928. Canna Crosses II. Meded. Landb. Hoogeschool, Wage-
ningen. Vol. 32. Paper I.

Huskins, C. Leon., and Stanly G. Smith, 1935. Meiotic chromosome
structure in TrilUum erectum L. Ann. of Bot., Vol. XLIX.

Koernicke, M., 1903. Der heutige Stand der pflanzlichen Zellforschung. Ber
d. d. bot. Ges. Bd. 21.

La Cour, L. F., 1931. Improvements in everyday technique in plant cytolo-
gy. Journ. Roy. Micr. Soc. 51.

Lawrence, W. J. C., 1931. The secondary association of chromosomes. Cy-
tologia. 2.

Leliveld, J. A., 1931. Some remarks on the cytology of Oenothera. Ree.
trav. bot. nöerl. Vol. 25.

Linsbauer, K., 1921 — 1934. Handb. der Pflanzenanatomie. Band I, Band
II, Band X 2.

De Meyere, j. C. H., 1928. Über haltbare, rasche Färbung vermittels Azeto-
Karmin. Zeits. f. Wiss. Mikr. u. f. mikr. Technik Bd. 46.

De Meyere, J. C. H., 1930. Über einige europäische Insekten, besonders
günstig zum Studium der Reif ungsteilungen, nebst einigen Zusätzen zur
Azeto-karmin Methode. Zool. Anz. Bd. 88.

Tokugawa I., and Kuwada Y., 1924. Cytol. stud, on some garden-varieties
of Canna. Jap. Journ. of Bot. Bd. 2.

Wiegand, K. M,, 1900. The development of the embryosac in some mono-
cotyledonous plants. Bot. Gazette. Bd. 30.

De terminologie voor de beschrijving van de reductie-deeling is
niet meer in overeenstemming met de bekende feiten.

De beschouwingen van Sands bewijzen niet, dat bij Oenothera
telosyndese voorkomt (Amer. Journ. of Bot., Vol. 21, 1934).

Ten onrechte beschouwt Tschermak als bewezen, dat de matro-
kliene nakomelingen in kruisingen van de geslachten Pisum. Lens,
Vicia en Ervillea berusten op quot;hybridogene parthenogenesisquot; (Der
Züchter, Jahrg. 7, 1935). ^^

De chromosomen-morphologie heeft, voorloopig althans, zeer ge-
ringe waarde voor de systematiek.

De zg. glaciaal-relicten onder de hoogere planten van ons land
worden ten onrechte als zoodanig beschouwd.

De proeven van Van der Paauw, met betrekking tot de opname
van water onder omstandigheden die de transpiratie verhinderen,
bevestigen de juistheid van de theorie van Sen en Blackman
(Recueil d. trav. bot. néerl., Vol. 32, 1935).

Het onderzoek van Burgeff maakt het waarschijnhjk, dat de
gunstige invloed van suikers op de kieming van Orchideeën-zaden
in reincultuur gedeeltelijk berust op de aanwezigheid van oligo-
dynamisch werkzame stoffen (Ber. dts. bot. Ges., Bd. 52, 1934).

De grondige behandeling van de planten-morphologie is voor de
leerlingen van de lagere klassen van inrichtingen voor middelbaar
en gymnasiaal onderwijs van groote waarde.

Een gepaste afwisseling van klassikaal onderwijs en individueele
zelfwerkzaamheid is voor het onderwijs in de biologie het meest
gewenscht.

Het quot;schoonheids-elementquot; in de levende natuur behoort bij het
onderwijs niet te worden vergeten.

Het zg. hospiteeren is voor den a.s. leeraar de belangrijkste voor-
bereiding voor het zelfstandig lesgeven.

äi, IJ-