/. Embryo/, exp. Morph. Vol. 33, 2, pp. 487-498, 1975
Printed in Great Britain
487
Development and differentiation
in vitro of Drosophila imaginal disc cells from
dissociated early embryos
By ANDREAS Dt)BENDORFER 1 , GLEN SHIELDS 2 AND
JAMES H. SANG2
From School of Biological Sciences, University of Sussex
SUMMARY
Embryos of Drosophila melanogaster, 6-8 h after oviposition, were dissociated and the cells
cultured in vitro. Besides larval cell types, imaginal disc cells, assembled and growing in
bloated monolayered vesicles, were obtained.
The cells of these vesicles become competent to differentiate adult structures when treated
with a-ecdysone or ecdysterone in vitro. Recognizable patterns of the adultflyare not formed
though. If metamorphosis of imaginal cell vesicles from in v/Vro-cultures is induced in vivo by
transplantation into host larvae of various ages within the third larval instar, recognizable
patterns can differentiate provided the host larva does not metamorphose prior to 2 days after
transplantation. The frequency of specific patterns in the implants can be increased by providing 9 days of culture in vivo (adult host flies) before metamorphosis. Passage through the
third larval instar is not essential for these cells to produce identifiable patterns since culture
in adult flies alone can achieve this. The quality of the differentiated pattern is not correlated
with the extent of cell proliferation in the cultured tissues.
The problem of pattern specification in vitro and in vivo is discussed.
INTRODUCTION
Hadorn and colleagues investigating the time of determination of imaginal
disc cells in the embryo of Drosophila melanogaster (Hadorn et ah 1968;
Schubiger, Schubiger-Staub & Hadorn, 1969) found that aggregates of dissociated tissue from embryos 10 ± 1 h old, when cultured in adult flies and then
put through metamorphosis in larvae, differentiated recognizable adult cuticular
structures. The structures correlated with the part of the embryo used as source
material, and it was concluded that imaginal disc cells are determined for at
least the general commitment 'anterior' and 'posterior' by the end of gastrulation. Using a slightly modified technique, Chan & Gehring (1971) showed that
the cells are so determined as early as just after blastoderm formation, at about
1
Author's address: Zoologisches Institut der Universitat Zurich, Kiinstlergasse 16, 8006
Zurich, Switzerland.
2
Authors'1 address: School of Biological Sciences, University of Sussex, Falmer, Brighton,
BN1 9QG.
488
A. DUBENDORFER, G. SHIELDS AND J. H. SANG
3 h after oviposition. The same conclusion was drawn by Illmensee and
Mahowald (1974) using heterotopic cell transplantation in embryos of the
blastoderm stage.
In most of these experiments (Hadorn et al. 1968; Schubiger et al. 1969; Chan
& Gehring, 1971) the embryonic material was cultured for 10-14 days in the
adult host prior to the metamorphosis test. During this period the imaginal disc
cells acquired their competence for metamorphosis. In situ this process is completed about 3 days after oviposition (Bodenstein, 1939a,b; Gateff, 1971; Mindek,
1972; Mindek & Nothiger, 1973). Besides the imaginal disc cells, various larval
tissues developed during the in vivo culture period. For the imaginal disc tissue
it is not known whether the ultimate specification, leading to the capacity to
form organized patterns within a disc, occurs autonomously (i.e. whether the
cultured tissue is self-organizing), or if it depends on the relative position of the
cells to other larval structures and/or on specific factors provided by them. The
aim of the present report is to investigate this question.
Cells of dissociated early Drosophila embryos can be cultured in vitro (reviewed in Shields & Sang, 1970) and we have found (Shields, Diibendorfer &
Sang, 1975) that in such cultures imaginal disc cells multiply, giving rise to
hollow, monolayered vesicles which can differentiate adult cuticle with bristles
and trichomes when transplanted into a metamorphosing larva. This has enabled us to study the processes of acquisition of competence to metamorphose
and of specification for adult structures in vitro, and to make a comparison with
in vivo findings.
Growth in vitro of imaginal disc cells, as monolayered vesicles, has been
described earlier by Schneider (1972), but her results cannot answer the question
of autonomy or otherwise of imaginal disc specification, as she used embryos of
an extremely late stage (shortly before hatching of the larva). Thus, the multiplying cells in her cultures had grown out from already detectable imaginal disc
primordia and had been in continuous contact with an embryonic fragment
containing organized larval tissues.
MATERIAL AND METHODS
Embryonic material for both in vitro and in vivo culture purposes was obtained from embryos of Drosophila melanogaster (Oregon) 6-8 h old. In vitro
cultures were established and maintained according to the procedures we have
previously described (Shields & Sang, 1970; Shields et al. 1975). Large-scale
cultures were also set up to obtain larger quantities of imaginal cell vesicles.
These involved 1 ml volumes of cell suspension in plastic Petri dishes (Nunclon,
30 mm diam.) which were sealed to prevent loss of CO2.
In vivo implantations of primary embryonic material and of in vitro cultured
cells were carried out according to the standard techniques of Ephrussi &
Drosophila imaginal disc cells in vitro
489
Table 1. Imaginal differentiation of reaggregated embryonic cells
after varying culture time in vivo
Days of culture
Type of differentiation*
metamorphosis
Total of analysed
implants
I
II
III
5
9
12
16
20
35
30
52
46
30
33
19
28
32
19
2
5
14
7
7
0
6
10
7
4
hpfo
A
* Type of differentiation: I, Cuticular spheres, plain or with some bristles and/or trichomes;
IT, Large structures with unidentifiable 'simple patterns', cf. p. 493; III, Differentiated
structures identified as a specific part of the adult.
Beadle (1936) and Hadorn (1963). The adult and larval hosts for these were
from the same stock as the donor embryos: genetic marking of the host animals
was not considered necessary as earlier studies (Hadorn, 1966; Gehring, 1966)
have shown that host cells do not contribute to the developing implant. Confirmation of this was obtained in our transplantation series in the cases where
genital structures of one sex were differentiated in an implant carried by a host
of the other sex. Host animals were maintained on standard food (cornmeal,
sugar, dry and live yeast) at 25 °C. The age of the larval hosts is given in hours
after oviposition.
The a- and /?-ecdysones used to induce differentiation in vitro in certain experiments were dissolved directly in the culture medium and diluted to the required
concentration. They were applied by performing medium changes on the cultures in the normal way, using ecdysone-containing medium. The a-ecdysone
was a generous gift from Schering AG and the /?-ecdysone was purchased from
Rohto Pharmaceutical, Japan. Fuller technical details for particular experiments are given where necessary in the results.
RESULTS AND CONCLUSIONS
Acquisition of competence and specification in vivo
To begin with, the effect of the dissociation procedure on the acquisition of
competence and specification in v/vowas tested. Immediately after preparation
of an embryonic cell suspension (as normally done for initiating cultures in
vitro) the cells were aggregated by centrifugation (1 min, 600 g) and pieces of the
aggregate (0-001-0-0045 mm3) implanted into adult females 1 day old. Batches
of 24 host flies were then dissected after 4, 8, 11, 15 and 19 days, respectively.
Normally, the developed implants contained differentiated larval tissues, often
actively pulsating, and lumps of undifferentiated imaginal disc tissue. The
490
A. DUBENDORFER, G. SHIELDS AND J. H. SANG
implants were cut into pieces and tested for their capacity to form adult fly structures by transplanting into larvae 100 ± 4 h oJd which metamorphosed about
24 h later. Thus, the overall culture time in vivo in the five series was 5, 9, 12, 16
and 20 days, respectively.
Differentiated cuticular structures were found in all five of the transplantation
series, and Table 1 details the types of differentiation obtained. With a 5-day
culture period (i.e. a time similar to that required in situ for the complete
development of the imaginal discs), spheres of cuticle containing bristles and
trichomes were found, but no identifiable specific adult structures. Specific
structures differentiated only after an in v/vo-culture period of 9 days or longer,
then appearing with frequencies varying from 13 to 20 % of the total differentiated structures. Derivatives of the following imaginal discs were obtained:
male and female genital disc (including internal parts), labial-, eye-antennal-,
wing-, haltere-, and leg disc.
We conclude that the state of competence for metamorphosis is achieved in
aggregates of dissociated embryonic tissue within 5 days or less with culture in
vivo, but that specification for final adult structures is completed only after a
culture period that clearly exceeds the time needed in situ. Thus, our experimental conditions delay the normal development of imaginal disc cells in vivo.
Acquisition of competence and specification in vitro
When cells of dissociated 6-8 h embryos are cultured in vitro, a variety of
larval cell types (nerve, muscle, fat-body, epidermal, etc.) differentiate during
the first 10 days of culture (Shields et al. 1975). The first imaginal disc cell
vesicles become apparent in the second or third week, in forms closeJy similar
to those described by Schneider (1972). New vesicles may arise in the cultures
continuously over a period of up to 3 months, and single vesicles can grow to
eventually contain many thousands of cells (Fig. 1 A, B).
It was discovered that the vesicles respond to ecdysone, and their competence
for metamorphosis was looked for directly in vitro by applying ecdysone to
cultures containing well-developed vesicles. Metamorphosis in vitro could be
induced with both a- and ^-ecdysone, and /?-ecdysone was found to be most
effective at a concentration 20-100 times lower (0-5-0-1 /tg/ml) than that of
a-ecdysone (10 /tg/ml).
A high proportion of the vesicles treated with ecdysone differentiated cuticle,
plain or with trichomes, normally on their inner surface. Usually the cuticle
remained unpigmented and the differentiated structures rather difficult to see,
even under a phase-contrast microscope; but pigmented cuticular structures were
also observed. Several different types of trichomes were noted, whereby one
vesicle always contained only one type. Bristles with fully developed socket and
shaft appeared in a number of the vesicles, with or without trichomes, and again
were of varied character (Figs. 1C, 2 A). But although in some vesicles up to 20
and in one case some 100 bristles were formed, the differentiated cuticular
Drosophila imaginal disc cells in vitro
491
Fig. 1. (A) Imaginal cell vesicle (V) in the third week of culture in vitro. (B) Part
(about one-twelfth) of a large grown vesicle, tenth week of culture (view on surface
of vesicle). (C) Differentiation in vitro of bristles (B) and trichomes (T) in an imaginal
cell vesicle treated with 5-0 /Mg/ml a-ecdysone.
492
A. DUBENDORFER, G. SHIELDS AND J. H. SANG
A
L
n in
i ii in
i II in
2 days
9 days
5-8 K
Fig. 2. Culture and metamorphosis of imaginal cell vesicles in vitro and in vivo.
Experimental design and summary of the results. Types of differentiation: I, cuticular
spheres, plain or with a number of trichomes and/or a few bristles; II, 'simple
pattern', cf. p. 493; III, differentiated pattern, identified as a specific part of the
adult. Times given in hours indicate the age after oviposition of the donor embryos
and of the host larvae, respectively, those given in days refer to the culture period.
M, metamorphosis; n, total of differentiated implants in the series B-F; PF, puparium
formation.
Drosophila imaginal disc cells in vitro
493
spheres never contained any disc-specific marker structures, such as eye facets,
claws, sensillae, etc., and we have therefore never been able to identify the type
of differentiation as being characteristic for a particular imaginal disc.
Inadequacy of the in vitro conditions to support full metamorphosis had to be
considered as a possible cause for the complete absence of specialized structures
in the differentiated vesicles. We therefore also tested the differentiative capacity
of in vitro grown vesicles in vivo, by transplanting them into larvae. Early (7274 h), middle (100 ± 5 h), and late third (immobile stage) larvae were used as
hosts, and in two of the series of experiments, an intermediate culture period
in young adult females preceded the implantation into the larva. Fig. 2 illustrates
the experimental design and the results.
From the transplantation of vesicles into immobile larvae, ready to pupate
(Fig. 2B), 65 differentiated implants were recovered. Most of these were identical to those obtained following treatment with ecdysone in vitro, i.e. small, with
or without trichomes, and/or a few bristles. The trichomes and bristles were the
same types as found in vitro, and again individual vesicles were always restricted
to only one type (although where clusters of vesicles were implanted, several
cuticular spheres with different trichome types were sometimes present in one
implant).
Among the bristle-bearing implants there were none containing specialized
structures allowing an unambiguous identification, even when the culture time
in vitro had been as long as 48 days. Many of the implants (15 %), however,
were large, and displayed a characteristic pattern, with a number of slim bristles
(often dozens) scattered over a large sheet of semi-transparent chitin which was
also fairly homogenously covered with thousands of trichomes (Fig. 3 A). This
pattern was similar to that normally developing from abdominal histoblasts, but
the implants could not be definitely identified as abdominal structures, since
neither the characteristic pattern of tergite pigmentation nor the typical distribution of bristles and trichomes were observed. When the differentiated
implants were classified, this type of cuticular structure - large, but not recognizable as belonging to a specific body region - was considered as a separate
class (cf. Fig. 2, class II and Table 1, II), and referred to as 'simple pattern'
(Fig. 3A).
We conclude from the results of differentiation of vesicles in vitro and in
metamorphosing larvae that: (i) acquisition of competence for metamorphosis
is achieved by imaginal disc cells in vitro, and hence does not depend on their
embryonic organization into disc primordia or on external influences from a
living host, (ii) specialized, complete adult structures do not develop from
embryonic imaginal cells grown in vitro when they are exposed directly to
metamorphosis hormones, either in vitro or in vivo.
In our further transplantations of vesicles into larvae of the mid and early
third instar, the differentiation response of those that were implanted one day
before metamorphosis (host: 100 ± 5 h, Fig. 2C) was much the same as for the
494
A. DUBENDORFER, G. SHIELDS AND J. H. SANG
rf .--%
A
Drosophila imaginal disc cells in vitro
495
late larval series; but one implant, out of 106, developed clearly identifiable
labial structures, and another one some sensillae, possibly of the leg. When the
implants had 2 days in the larva prior to metamorphosis (host: 72-74 h, Fig.
2D), 8 % differentiated recognizable adult structures, such as genital apparatus,
antennal parts with aristae and typical sensillae, leg parts, labial structures, etc.
(cf. Fig. 2, class III, and Fig. 3B).
The proportion of recognizable implants could be increased even more (up
to 40 %) if in vitro grown tissue was cultured in an adult fly for 9 days before
being submitted to metamorphosis in a larva (Fig. 2E). The percentage of
identifiable adult structures (class HI in Fig. 2E, F) is about the same in old and
young larval hosts. This shows that growth in the third larval instar is not vital
for the specification of recognizable adult structures, as culture in adult flies
alone gives the same result.
Implants with recognizable structures often contained all the parts normally
derived from one imaginal disc, organized correctly as in the normal fly (Fig.
3 B). In the case of genital structures, however, it was found that anal plates were
always missing from the inventory, occurring as separate structures (Fig. 3D, E).
This suggests that the larval genital disc is the product of fused embryonic
primordia from more than one abdominal segment, the fusion occurring after
gastrulation. Such is known to be the case in other Dipteran flies, e.g. Musca
domestica, where the fusion occurs only during metamorphosis (Diibendorfer,
1970, 1971;Emmert, 1972).
Occasionally structures belonging to two or more different imaginal discs were
identified in a single implant. We cannot say whether these differentiated as a
result of transdetermination occurring in vitro (Hadorn, 1966; Gehring, 1972),
as with our technique all implants also had a short culture period in vivo prior
to metamorphosis. Also, parts of more than one vesicle might occasionally have
been transplanted together, and we have no means of identifying the type of
determination of the vesicles before metamorphosis.
DISCUSSION
Earlier reports (review by Nothiger, 1972) suggest that the imaginal disc cells
of the embryonic stage (6-8 h) used to initiate our cultures are already determined for their general disc properties. Such cells multiply in vitro to form the
Fig. 3. (A) 'Simple pattern' differentiated by an in vitro grown imaginal cell
vesicle after metamorphosis in vivo (host: late third instar larva). (B) Full complement of labial structures differentiated from an in vitro grown imaginal cell vesicle
after an additional culture period of 9 days in vivo. Lk, 'Labellarkalotte'; Pm,
prementum; Pt, pseudotracheae. (C) Very small implant carrying recognizable leg
bristles, accompanied by bracts (Br). (D) Isolated female vaginal plate (Vp) with
'longbristles' (£). (The tiny bristles of the 8th tergite are present but not visible
on the photograph.) (E) Isolated female anal plate (An).
496
A. DUBENDORFER, G. SHIELDS AND J. H. SANG
vesicles which we find (Fig. 1 A, B). The crucial question is why such vesicles do
not become competent to form particular complements of adult structures (leg,
wing, etc.) in vitro, or upon immediate metamorphosis in vivo, but, instead, require a prolonged period of growth in either a larval or an adult environment
before they differentiate fully. Nevertheless, vesicles are capable of laying down
chitin when exposed to ecdysone, and this chitin may contain 'simple patterns'
of bristles and trichomes (p. 493). Schneider (1972) found that vesicles growing
out from late embryos cultivated in vitro behaved similarly, and it is interesting
that in this case, too, specialized structures such as aristae, sensillae, mouth
parts or bracts (leg) were never found. It is unlikely that this was due to the small
sample size (31 implants) since similar samples in our experiments produced
these advanced structures under appropriate culture conditions. So even if
vesicles of disc cells derive from nearly hatched embryos, and remain attached
to their site of origin, they are unable to achieve an advanced degree of specification when cultured in vitro.
Cell divisions have been reported to be a necessary prerequisite for cell
differentiation and pattern formation in many developmental systems (cf.
Ursprung, 1968). For imaginal discs cell divisions seem to play an important
role in the acquisition of competence for metamorphosis (Mindek & Nothiger,
1973) as well as for transdetermination (Gehring, 1966; Tobler, 1966; Wildermuth, 1968; Mindek, 1968). However, Lee & Gerhart (1973) found that transdetermination is not solely dependent on cell divisions but also on some other
(yet unknown) factors. In our system disc cell growth is apparently slow in
culture, but ultimate vesicle size may be as large as that of full-grown discs. It is
therefore unlikely that full maturation is solely dependent on cell divisions. In
fact, after culture in vivo, we obtained easily identifiable small implants (Fig. 3 C)
where the number of cell divisions must have been limited. Thus, the successful
and complete differentiation of imaginal discs appears to depend on the constitution of the larval or adult haemolymph, as well as on cell division.
If the ultimate organization of an imaginal disc is the result of a continuous
determinative process, promoted by and solely dependent on cell divisions, we
should expect a whole range of partial patterns from cultured embryonic
imaginal material, irrespective of whether the cells have divided in vitro or in
vivo. But this has not been found. Certainly such 'organ districts' or 'regional
blastemas' (see Bryant, 1974, for discussion) responding to ecdysone have been
identified among disc transplants and regenerates, but these derive from already
determined tissues and may be dependent on organizational relationships within
the material, as Postlethwait & Schneiderman (1973) suggest. If this is so, failure
to attain this state could be a consequence of the vesicular growth pattern itself.
The normal growth of disc cells as a columnar epithelium, which becomes a
folded sac, may be the simple, but essential, form for meaningful intra-disc
specification through cellular interactions. This form is achieved by vesicles of
imaginal cells that are cultured in vivo for some time. Alternatively and/or
Drosophila imagined disc cells in vitro
497
additionally, it may be a consequence of the different chemistry of the haemolymph, compared with that of the synthetic culture medium, and an improved
medium may thus permit normal disc development.
ZUSAMMENFASSUNG
Drosophila melanogaster - Embryonen wurden im Gastrula Stadium (6-8 Stunden nach
Eiablage) dissoziiert und die Zellen in vitro kultiviert.
Neben larvalen Zelltypen bringen solche Kulturen auch aufgeblahte, einschichtige Blasen
prospektiver Imaginalscheibenzellen hervor. Diese konnen mit a-Ecdyson oder Ecdysteron
zur Metamorphose in vitro veranlasst werden. Spezifische Muster der Adultcuticula, wie z.B.
Augenfacetten, Klauenorgane oder Sensillen, werden nicht differenziert in vitro, sondern
lediglich einfache Borsten- und Trichommuster ('simple patterns').
Wenn in vitro entstandene Imaginalzell-Blaschen nach Transplantation in Wirte des dritten
Larvenstadiums in vivo metamorphosieren, konnen normal organisierte Adultmuster differenziert werden. Solche Leistungen sind aber auf jene Serien beschrankt, bei welchen die
Metamorphose erst 2 Tage nach der Transplantation erfolgt. Ein Aufenthalt von 9 Tagen in
transitorischen Adultwirten vor dem Metamorphosetest steigert die Frequenz identifizierbarer
Muster in den Implantaten, auch wenn das dritte Larvenstadium dabei umgangen wird. Es
wurde keine Korrelation zwischen der Art der Musterbildung und dem Ausmass der Zellproliferation gefunden.
Das Problem der Musterbildung in vitro und in vivo wird diskutiert.
This work was supported by grants from the Science Research Council of Great Britain.
A.D. had a fellowship of the European Exchange Programme. We thank Dr R. Nothiger for
commenting on the manuscript.
REFERENCES
BODENSTEIN, D. (1939a). Investigations on the problem of metamorphosis-V. Some factors
determining the facet number in the Drosophila mutant Bar. Genetics, Princeton 24, 494508.
BODENSTEIN, D. (19396). Investigations on the problem of metamorphosis. IV. Developmental relations of interspecific organ transplants in Drosophila. J. exp. Zool. 82, 1-30.
BRYANT, P. J. (1974). Determination and pattern formation in theimaginal discs of Drosophila.
In Curr. Top. Devi Biol. (ed. Ruddle). New York: Academic Press.
CHAN, L.-N. & GEHRING, W. (1971). Determination of blastoderm cells in Drosophila
melanogaster. Proc. natn. Acad. Sci. U.S.A. 68, 2217-2221.
DUBENDORFER, A. (1970). Entwicklungsleistungen transplantierter Genital- und Analanlagen von Musca domestica und Phormia regina. Experientia 26, 1158-1160.
DUBENDORFER, A. (1971). Untersuchungen zum Anlageplan und Determinationszustand der
weiblichen Genital- und Analprimordien von Musca domestica L. Wilhelm Roux Arch.
EntwMech. Org. 168, 142-168.
EMMERT, W. (1972). Entwicklungsleistungen abdominaler Imaginalscheiben von Calliphora
erythrocephala (Insecta, Diptera). Experimented Untersuchungen zur Morphologie des
Abdomens. Wilhelm Roux Arch. EntwMech. Org. 169, 87-133.
EPHRUSSI, B. & BEADLE, G. W. (1936). A technique of transplantation for Drosophila. Am.
Nat. 70, 218-225.
GATEFF, E. (1971). Developmental capacities of immature eye-antennal discs of Drosophila.
Ph.D. thesis, University of California, Irvine, U.S.A.
W. (1966). Ubertragung und Anderung der Determinationsqualitaten in Anten-
GEHRING,
nenscheiben-Kulturen von Drosophila melanogaster. J. Embryol. exp. Morph. 15, 77—111.
498
A. DUBENDORFER, G. SHIELDS AND J. H. SANG
W. (1972). The stability of the determined state in cultures of imaginal discs in
Drosophila. In The Biology of Imaginal Discs (ed. H. Ursprung & R. Nothiger), pp. 35-58,
New York: Springer Verlag.
HADORN, E. (1963). Differenzierungsleistungen wiederholt fragmentierter Teilstiicke mannlicher Genitalscheiben von Drosophila melanogaster nach Kulter in vivo. DevlBiol. 7, 617—
629.
HADORN, E. (1966). Konstanz, Wechsel und Typus der Determination und Differenzierung
in Zellen aus mannlichen Genitalanlagen von Drosophila melanogaster nach Kultur in vivo.
DevlBiol.l, 617-629.
GEHRING,
HADORN, E., HURLIMANN, R., MINDEK, G., SCHUBIGER, G. & STAUB, M. (1968). Entwick-
lungsleistungen embryonaler Blasteme von Drosophila nach Kultur im Adultwirt. Revue
suisse Zool. 75, 557-569.
ILLMENSEE, K. & MAHOWALD, A. P. (1974). Transplantation of posterior polar plasm in
Drosophila. Induction of germ cells at the anterior pole of the egg. Proc. natn. Acad. Sci.
U.S.A. 11, 1016-1020.
LEE, L.-W. & GERHART, J. C. (1973). Dependence of transdetermination frequency on the
developmental stage of cultured imaginal discs of Drosophila melanogaster. Devi Biol. 35,
62-82.
MINDEK, G. (1968). Proliferations- und Transdeterminationsleistungen der weiblichen
Genital- lmaginalscheiben von Drosophila melanogaster nach Kultur in vivo. Wilhelm
Roux Arch. EntwMech. Org. 161, 249-280.
MINDEK, G. (1972). Metamorphosis of imaginal discs of Drosophila melanogaster. Wilhelm
Roux Arch. EntwMech. Org. 169, 353-356.
MINDEK, G. & NOTHIGER, R. (1973). Parameters influencing the acquisition of competence
for metamorphosis in imaginal discs of Drosophila. J. Insect Physio!. 19, 1711-1720.
NOTHIGER, R. (1972). The larval development of imaginal discs. In The Biology of Imaginal
Discs (ed. H. Ursprung & R. Nothiger), pp. 1-34. New York: Springer Verlag.
POSTLETHWAIT, J. H. & SCHNEIDERMAN, H. A. (1973). Developmental genetics of Drosophila
imaginal discs. A. Rev. Genetics 7, 381-434.
SCHNEIDER, I. (1972). Cell lines derived from late embryonic stages of Drosophila melanogaster. J. Embryol. exp. Morph. 27, 353-365.
SCHUBIGER, G., SCHUBIGER-STAUB, M. & HADORN, E. (1969). Mischungsversuche mit Keimteilen von Drosophila melanogaster zur Ermittlung des Determinationszustandes imaginaler
Blasteme im Embryo. Wilhelm Roux Arch. EntwMech. Org. 163, 33-39.
SHIELDS, G., DUBENDORFER, A. & SANG, J. H. (1975). Differentiation in vitro of larval cell
types from early embryonic cells of Drosophila melanogaster. J. Embryol. exp. Morph. 33,
159-175.
SHIELDS, G. & SANG, J. H. (1970). Characteristics offivecell types appearing during in vitro
culture of embryonic material from Drosophila melanogaster. J. Embryol. exp. Morph. 23,
53-69.
TOBLER, H. (1966). Zellspezifische Determination und Beziehung zwischen Proliferation und
Transdetermination in Bein- und Fliigelprimordien von Drosophila melanogaster. J.
Embryol. exp. Morph. 16, 609-633.
URSPRUNG, H. (1968). The Stability of the Differentiated State. New York: Springer Verlag.
WILDERMUTH, H. (1968). Differenzierungsleistungen, Mustergliederung und Transdeterminationsmechanismen in hetero- und homoplastischen Transplantaten der Riisselprimordien
von Drosophila. Wilhelm Roux Arch. EntwMech. Org. 160, 41-75.
{Received 14 August 1974, revised 18 September 1974)