Development 112, 755-761 (1991)Printed in Great Britain © The Company of Biologists Limited 1991

755

The embryonic origin of imaginal discs in Drosophila

MICHAEL BATE and ALFONSO MARTINEZ ARIAS

Department of Zoology, Downing Street, Cambridge, CB2 3EJ, UK

Summary

The thoracic imaginal discs of Drosophila melanogastercan be observed during embryogenesis as clusters of cellswith particular shapes, sizes and behaviours. Thesestructures can be detected soon after germ band

shortening and their development appears to be tightlylinked to that of the larval epidermis.

Key words: imaginal discs, embryos, Drosophila.

Introduction

The adult epidermis in Drosophila develops fromprecursor cells that are set aside during embryogenesis.While the abdominal segments are derived fromhistoblasts, clusters of cells that contribute to the larvalepidermis and proliferate in the pupa, the rest of theadult epidermis is formed from the imaginal discs,groups of cells that do not contribute to the larvalepidermis and proliferate during larval life (Nothiger,1972).

It is known that the segregation of the imaginal discprecursors from the larval epidermis takes place duringembryogenesis, and the imaginal discs can be identifiedin the first instar larva (Madhavan and Schneiderman,1978). However, the embryonic events of segregation,as well as the number and position of the precursors ineach primordium, have never been demonstrateddirectly, but inferred from the results of a variety ofdifferent experiments. Data from clonal analysis (Wies-chaus and Gehring, 1976a) and gynandromorph (GarciaBellido and Merriam, 1969; Wieschaus and Gehring,1976b) or ablation fate mapping (Lohs Schardin et al.1979) suggest that soon after cellular blastoderm,ectodermal cells are selected to become imaginalprecursors and thus become determined as differentfrom the larval cells.

Poulson (1950) observed some of the imaginal discs inthe Drosophila embryo after cuticle deposition and thefull complement has been described in the first instarlarva shortly after hatching, when the imaginal cellsstart to proliferate (Madhavan and Schneiderman,1978). Recently, the location of the primordia for theleg discs has been inferred from the pattern ofexpression of the gene Distalless (Dll) during earlyembryogenesis (Cohen, 1990). In the extended germband, small patches of Dll expression on the lateralsides of the thoracic epidermis have been postulated tobe the primordia of the leg discs (Cohen, 1990). Thissuggestion is difficult to confirm because there is, as yet,

no independent description of the segregation andposition of the imaginal primordia. The only relevantobservation is the description of the segregation of thethoracic discs from the larval epidermis in embryos ofthe fly Dacus tryoni at lateral positions in the thoracicepidermis (Anderson, 1963).

In this paper, we describe the embryonic origin of thethoracic imaginal discs in Drosophila and place it in thecontext of the patterning of the larval epidermis. Wehave used flat preparations and antibody staining ofDrosophila embryos at different stages of developmentto locate imaginal disc precursors and to follow theirsegregation from the larval epidermis. In addition,BUdR labelling at different stages of embryogenesissuggests that the definition and segregation of theseprimordia in the thorax are linked to the proliferation ofthe larval epidermis.

Materials and methods

StocksIn these experiments we have used OregonR and Canton S aswild-type stocks. To produce mutants for genes of thebithorax complex, we have used two stocks: DpP5;Df(3R)P9, which produces embryos lacking the whole of thebithorax complex (Lewis, 1978) and l/fcx922/TMl, whichproduces embryos mutant for the Ultrabithorax gene, Ubx.

BUdR incorporation, antibody stains andhistochemical stainsThe incorporation and visualization of BUdR followedpublished protocols (Bodmer et al. 1989) with minorvariations. The antibody DA.1B6 (Brower et al. 1980) alsoknown as anti-fasciclin III (Patel et al. 1987) was kindlyprovided by M. Wilcox. All manipulations of embryos for andduring antibody stains followed standard protocols (seeAshburner, 1990). Flat preparations of embryos, dissectionsand toluidine blue staining have been described before (Bate,1990).

756 M. Bate and A. Martinez Arias

Results

The larval epidermis and the imaginal discs afterembryonic proliferationBetween 12 and 14 h of embryogenesis, shortly beforethe onset of cuticle deposition, the larval epidermisconsists of an ordered array of epithelial cells of about10-12 cells long and 30 cells wide per hemisegment. Inflat preparations stained with toluidine blue, cellsappear slightly elongated in the dorsoventral axis andarranged in parallel lines at right angles to theanteroposterior axis (Fig. IB). The epidermal cellshave large nuclei and display different shapes and sizesat different levels of each segment. The overall featuresof a segment, in terms of number and arrangement ofcells, are similar in the thoracic and the abdominalsegments.

We have not succeeded in identifying the abdominalhistoblasts, but, at this stage, dorsal (wing and haltere)and ventral (leg) thoracic discs are easily visible in flatpreparations, or in wholemount embryos stained withthe monoclonal antibody DA.1B6 (Brower et al. 1980),also known as anti-fasciclin III (Patel et al. 1987). Theimaginal discs can be seen as subepidermal clusters ofcells still connected to the outside layer of larvalepidermis (Figs 1B,C and 3B). In Tl, there is aprominent invagination associated with the anteriorspiracle; this invagination is missing in T2 and T3 wherein a more ventral and posterior position the wing andhaltere discs can be found attached to the trachealsystem (Fig. 2). The wing disc contains about 24 cells,has a fan-like appearance and is located posteriorly inthe segment, hanging over T3. The haltere disc containsabout 12 cells, lies in a homologous position in T3 andforms a smaller, rounder structure than the wing disc.The cells of these discs are highly elongated and theirextended processes terminate as coherent groups in theepidermis of T2 and T3, where they form conspicuousclusters of end feet marking the sites of invaginationamong the larval epidermal cells (Figs 1 and 3). Theseclusters lie 3-4 cells anterior to the segment border(defined as the line of cells on which the larvallongitudinal muscles insert) and at the level of theremnant of the tracheal pit in the dorsoventral axis. Theleg discs appear as prominent spherical clusters oneither side of the CNS towards the middle of eachthoracic segment with the prothoracic discs located in amore medial position than the other two pairs(Fig. 1B,C). Unlike the wing and haltere discs, the legdiscs remain close to the overlying epidermis, formingcompact clusters of cells beneath the depressions thatmark the sites of invagination and of the future Keilin'sorgans. In embryos treated with anti-fasciclin, thecentre of each depression is marked by intense staining(Figs 1A and 4). We take it that this stainingdemarcates the anlagen of the Keilin's organs. Anestimate of the number of epidermal precursors in theleg discs is difficult because, unlike the dorsal discs,these represent a mixed population of epidermal, nerveand mesodermal cells (Bate et al. 1991). We areconfident that these groups of invaginated cells corre-

Fig. 1. Embryos stained for fasciclin III expression (A)and with toluidine blue (B and C) to show the dispositionof the disc anlagen at 9-10h (A) and 12-13h (B,C) afteregg laying. (A) Expression of fasciclin III in the epidermisof a 9-10 h old embryo. Most of the epidermal cells arestained; the leg discs are prominent in the ventral thorax asregions with low levels of stain (arrows) and a high level atthe centre, which presumably identifies the primordia ofthe Keilin's organ. A clear outpocketing of the epidermis isvisible more dorsally and posteriorly in each thoracicsegment (arrowheads). In T2 and T3, these correspondwith the wing and haltere disc primordia as they start theircell shape changes. In Tl, the outpocketing lies ventral tothe anterior spiracle and might correspond to theprothoracic disc. These structures are not present in theabdominal segments. (B) Ventral epidermis of the thoracicregion of a 12-13 h old embryo, dissected flat and stainedwith toluidine blue. The sites of invagination of the leg (1),wing (w), haltere (h) discs and of the anterior spiracle(arrow) are clearly visible. (C) Detail of embryo illustratedin B, to show imaginal disc cells. The leg discs (1) areprominent in this plane of focus, as are the attenuated feetof the wing (w) and haltere (h) discs. Scale bars: (A)50jun; (B and C) 35/m\.

Fig. 2. Internal structures of the thorax of a 13 h oldembryo, dissected flat and stained with toluidine blue. Inthis plane of focus, the anterior spiracle (s), wing (w) andhaltere (h) discs are clearly visible attached to the trachealsystem. Ventral down, dorsal up, anterior to the left,segments as indicated. Scale bar: 25 ixm.

spond to the thoracic imaginal discs because of theirposition, size, segmental specificity, attachments to theepidermal derivatives and behaviour in mutant back-grounds (see below). In addition, we have followed thecell clusters into the first instar larvae, where theycoincide with the imaginal discs as described in earlierreports.

Segregation of imaginal discsIt is possible to trace back the imaginal discs to earlierembryonic stages. In the case of the wing and halterediscs, invagination is initiated at 9-10 h after egg laying(AEL) and is signalled by a local reduction in the apicaldiameter of cells and a basal location of their nuclei, asthe cells elongate inwards (Fig. 3). This identifiesgroups of prospective wing and haltere cells locatedposteriorly in their respective segments (Fig. 1A) and atthe level of the persistent remnants of the tracheal pits.We have not been able to identify the cells of the dorsaldiscs before this, and because there may have beenextensive cell movements, we cannot say with certaintythat the discs are derived from an earlier population ofcells located at precisely this position in the developingsegment. As for the leg discs, they too can be identifiedas invaginating cell groups in embryos 9-10 h AEL(Figs 3 and 4). However, they form part of a set of cellsthat clearly differs in its behaviour from the rest of theepidermis in a thorax-specific way from early on.Specifically, the cells of the prospective leg discs have arounded mobile appearance, which contrasts with themore compact and closely packed cells of the abdomen(see Fig. 3A). Prior to invagination, and as far back as

Imaginal disc primordla 757

Fig. 3. Segregation of imaginaldiscs from larval epidermis inembryos revealed by stainingflat preparations with toluidineblue. (A) Epidermis of T2, T3and Al in a 9-10h oldembryo. The cells of the wing(w) and haltere (h) discs arebeginning to reduce theirdiameter apically; this willeventually result in theattenuation of the end feetshown in B. The leg discs (1)are prominent evaginations ofthe ventral epidermis, with aprominent depression(arrowheads), where cells mayalready be invaginating.(B) After invagination, the legdiscs (1) at this plane of focuscan just be distinguished asclumps of cells underneath theepidermis. The dorsal discsremain in contact with theepidermal sheet by theirattenuated end feet (arrows).Segments as indicated. Scalebar: 20 fim.

the onset of germ band shortening, this cell group,which lies just dorsal to the CNS and posterior in thesegment, is conspicuous for its divergent cell shape andarrangement, properties that suggest that the cellsalready have distinctive adhesive characteristics, and/or that they are undergoing a specific programme ofproliferation. One interesting behaviour of the prospec-tive disc regions is a transient reduction in the levels offasciclin III expression just prior to invagination, whichcould signal a change in adhesive properties (Fig. 4).

Epidermal proliferation during embryogenesisAfter proliferation, the cells of the imaginal discsrepresent a substantial fraction of the embryonicepidermis (about 10 %, i.e. some 40 cells out of a totalof 300 cells per hemisegment). Assuming that duringembryogenesis all epidermal cells divide the same

number of times, the larval epidermis of the thoracicand abdominal segments after invagination of the disccells, ought to consist of different numbers of cells.However, cell counts from embryos after germ bandshortening and the segregation of the discs, indicatethat, within the limits of observation, thoracic andabdominal segments have a similar number of epider-mal cells. Since the separation of the imaginal cells fromthe larval epidermis takes place after the last epidermalmitosis, this suggests either that there are differentnumbers of mitoses in thoracic and abdominal seg-ments, or that the discs initiate from very smallprimordia, which proliferate quickly after invagination.

To distinguish between these alternatives, we haveused BUdR as a label to monitor DNA synthesis andproliferation in the embryonic epidermis (Truman andBate, 1988; Bodmer et al. 1989). When embryos are

758 M. Bate and A. Martinez Arias

"1

as*

B

Fig. 4. Epidermis of embryosstained for fasciclin IIIexpression. (A) 10-11 hembryo. The sites of the dorsaldiscs during their segregationfrom the larval epidermis canbe seen as a local reduction inlevels of fasciclin expression(arrowheads). The leg discshave a heavily stained centralregion corresponding to thelocation of the primordium forthe Keilin's organs (arrows).(B) After imagination, theexpression of fasciclin IIIdemarcates the endfeet of thewing (w) and haltere (h) discsand the position of the Keilin'sorgans (arrows). Scale bar:20 /jm.

continuously exposed to BUdR from 5 to 12 h AEL, allepidermal cells are labelled (Fig. 5A). If a similarexperiment is performed from 6h AEL onwards(Fig. 5B), conspicuous label appears preferentially inthe ventral region of the thoracic segments. In contrast,labelling from 7.5 h AEL shows no epidermal deriva-tives labelled other than the sense organs (Bodmer etal.1989, and unpublished observations) and some cells ofmesodermal origin associated with the imaginal discs.These results suggest that the extra proliferation in thethorax provides a pool of cells for the process ofsegregation.

Disc development in homoeotic mutantsThe segmental specificity of the imaginal discs and theirderivatives is under the control of the homoeotic genesand genetic evidence suggests that the assignation ofimaginal primordia to particular segments occurs duringembryogenesis (Morata and Kerridge, 1981). To con-firm this, we have looked for the presence or absence of

discs in embryos with mutations in Ubx, whichtransforms parasegment (PS) 5 and PS6 into PS4 andembryos homozygous for Df(3R)P9, which removes thecomplete bithorax complex and transforms PS5 throughPS13 into PS4 (Lewis, 1978). Embryos homozygous forthe Ubx922 allele have a set of extra leg discs, arudimentary wing disc and two new sets of spiracles(Fig. 6). Embryos homozygous for Df(3R)P9 lack thewhole bithorax complex and display a reiteration of apattern of dorsal and ventral discs between Tl and ' A8'.In these embryos, we identify dorsal disc primordia bythe characteristic position, shape and clustering of theirattenuated end feet, never seen in wild-type abdominalsegments. Using this criterion, we see dorsal discprimordia variably in segments from T2-'A7', althoughwe cannot say whether the primordia correspond towings or halteres. Leg discs and spiracles are moreobvious and are present in a reiterated pattern betweenTl and 'A8'. Posteriorly, it becomes progressively moredifficult to identify the end feet clusters and, in the more

Imaginal disc primordia 759

Fig. 5. Embryos treated with BUdR for different periods of time and stained with anti BUdR-antibody to reveal patternsof replication during the treatment. (A) Ventral view of an embryo that has been treated from 4.5-5 h AEL until 12-13 hAEL. All epidermal cells contain label, suggesting that during the period all cells replicated their DNA. The thoracic cells,however, show heavier labelling suggesting that they have undergone more replication and hence more proliferation.(B) Ventrolateral view of an embryo, which has been treated from 5.5-6h AEL until l l-12h AEL. The label ispreferentially ventral and there is a clear thoracic label which can be correlated with the prominently labelled region in A.Scale bar: 20 fan.

posterior segments, the number of end feet may be asfew as two or three, indicating that only a small numberof cells is taking part in the formation of individualprimordia.

Discussion

During the second half of embryogenesis, the thoracicimaginal discs are conspicuous, tightly packed subepi-dermal clusters of cells with denned positions andshapes. They begin to separate from the epidermalsheet through characteristic cell movements after germband shortening but remain connected to it by the endsof the invaginating cells. In the dorsal discs theseconnections become greatly elongated and attenuated.Our observations agree with those of Anderson (1963)for the discs of Dacus tryoni, in that the segregation ofthe discs from the larval epidermis takes place frompositions about half way down the dorsoventral axis.Although at the time of invagination, the primordia forthe ventral and dorsal thoracic discs are very close,about two cell diameters, we never see any connectionbetween them as has been described for Dacus.However, cells that we have elsewhere identified asmesodermal (Bate et al. 1991) do form an obviousstrand between them, once invaginated. The closespatial relationship between ventral and dorsal deriva-tives helps explain why it is that clones generated earlyin embryogenesis can cross between leg and wing or legand haltere (Wieschaus and Gehring, 1976a).

At about six and a half hours AEL, segment-specificpatterns of DNA replication can be observed thatprefigure the next round of divisions (unpublishedobservations). These patterns precede the segregationof the discs and highlight an extra division that willinvolve ventral thoracic cells, preferentially. Thiscellular behaviour may reflect some kind of pro-grammed compensatory division of epidermal cells forthe loss of the invaginating primordia. In this context, itis interesting that a major invagination event, whichoccurs at the site of the future anterior spiracle, is

consistently preceded by a similarly segment-specificand prominent pattern of DNA replication. Such anincrease is also visible in the ventral thorax and may beassociated with the development of the leg discs.However, it is clear that the invagination of the wingand haltere discs is not associated with any obviouslocalised round of DNA synthesis and proliferation. Weconclude that compensation for the loss of theinvaginated cells must involve changes in cell shape andcell arrangement and that the extra ventral divisions inthe thorax may compensate for the change in cellnumbers associated with the invagination of both thedorsal and ventral discs. We estimate (simply bycounting invaginating cells) that there are about 12 cellsin the haltere disc and 24 cells in the wing. Clearly thesecould be under- or overestimating the actual number ofcells in the disc. Counting cells in the first instar larva,Madhavan and Schneiderman (1978) found 37 cells inthe wing disc and 20 cells in the haltere. Either there hasbeen an additional round of cell division between thetime of our observations and theirs, or they includedcells in their estimate that we have shown (Bate et al.1991) to be of mesodermal origin. Our observations oncell number are difficult to relate to the estimates forthe disc anlagen obtained from clonal analysis (forreview see Wieschaus, 1978), because these refer to theprimordial cells in the blastoderm from which the discsare ultimately derived.

Clonal analysis has suggested that the imaginal discsare determined soon after the cellular blastoderm stage(Wieschaus, 1978). However, it is likely that at thisearly stage a population exists from which the specificimaginal discs will be drawn later in embryogenesis(Wieschaus and Gehring, 1976a,b). Our observationsthat specific cellular properties characteristics of disccells are not visible until half way through embryogen-esis agree with this suggestion. In addition, we believethat the development of the disc precursors is tightlyintegrated with the development of the pattern of thelarval epidermis as a whole. In hemimetabolous insects,the development of limbs is a continuous process

760 M. Bate and A. Martinez Arias

Fig. 6. Expression of fasciclinIII in embryos with mutationsin the bithorax complex.(A) Embryo homozygous forUbx922, snowing two extra setsof spiracles (arrowheads), andan extra set of leg and wingdiscs (arrows). (B) View of thedorsal epidermis of ahomozygous Df(3R)P9embryo, which can be easilyidentified by its failure toretract its germ band properly.The spiracles (dots) and thewing/haltere discs (arrows) canbe seen to be reiterated in the'abdominal region'.(C) Ventral view of the sameembryo showing the reiterationof the leg discs (arrows), whichcontinue in a different plane offocus. N.B. although leg discsare here identified by Keilinsorgans, there are clusters ofcells associated with them (notshown at this focal plane) thatcorrespond to leg primordia.Scale bar: 20 (im.

involving a sequence of determination, growth anddifferentiation. However, in holometabolous insectslike Drosophila, there is an additional step in thedevelopment of structures such as legs and wings in thatthese cells are 'adult' and this adult state must be

specified in some way so that the behaviour of thesecells becomes different from their larval neighbours.These adult cells remain in a quasi-undifferentiated,proliferative state and invaginate as discrete clusters,while their neighbours differentiate as a characterise-

Imaginal disc primordia 761

cally patterned population of epithelial cells. It is thesedistinctive properties of the future adult cells in theembryo (different shape, size and movement) that wehave described here and which have enabled us toidentify the embryonic precursors of adult structures.

The imaginal disc precursors represent embryonicstructures that appear as a response to the machinerythat establishes cell states and thus positional infor-mation in the embryo (Cohen, 1990; Simcox etal. 1989).These structures will proliferate during larval develop-ment but during early embryogenesis the precursors arepart of the epithelial sheet in which cell diversity isbeing generated. For this reason, we believe that theyparticipate in the generation of positional informationwithin the developing segment. Once this has beenachieved, and the segmental pattern is complete, thespecial adult characteristics of these cells becomemanifest, and trigger their withdrawal from the larvalepidermis. In our view, mutations that affect theestablishment of the disc primordia are likely to belethal mutations, which will also affect the overalldevelopment of pattern in the embryo. This wouldagree with the fact that despite many attempts to findmutations that specifically affect the definition of theimaginal discs as larval viable mutations, all that hasbeen found are mutations that affect the proliferation ofthe imaginal cells (Shearn and Garen, 1974; A. Shearnand A. Simcox, personal communication). Indeed, theelegant transplantation experiments of Simcox et al.(1990) show that some genes of the segment polarityclass, which are necessary for the proper patterning ofthe larval epidermis, are also required for properformation of imaginal disc precursors.

We are grateful to The Wellcome Trust for supportthroughout this work and to H. Skaer, A. Bejsovec and F.Gonzalez for reading the manuscript.

References

ANDERSON, D. T. (1963). The embryology of Dacus tryoni 2.Development of imaginal discs in the embryo. J. Embryol. exp.Morph. 11, 339-351.

ASHBURNER, M. (1990). Drosophila. A Laboratory Manual. ColdSpring Harbor, New York: Cold Spring Harbor.

BATE, M. (1990). The embryonic development of larval muscles inDrosophila. Development 110, 791-804.

BATE, M., RUSHTON, E. AND CURRIE, D. (1991). Cells withpersistent twist expression are the embryonic precursors of adultmuscles in Drosophila. Submitted to Development.

BODMER, R., CARRETTO, R. AND JAN, Y. N. (1989). Neurogenesis

of the peripheral nervous system in Drosophila embryos: DNAreplication patterns and cell lineages. Neuron 3, 21-32.

BROWER, D. L., SMITH, R. J. AND WILCOX, M. (1980). A

monoclonal antibody specific for diploid epithelial cells inDrosophila. Nature 285, 403-405.

COHEN, S. M. (1990). Specification of limb development in theDrosophila embryo by positional cues from segmentation genes.Nature 343, 173-177.

GARCIA BELLIDO, A. AND MERRIAM, J. (1969) Cell lineage of theimaginal discs in Drosophila gynandromorphs. J. exp. Zool. 170,61-76.

LEWIS, E. B. (1978). A gene complex controlling segmentation inDrosophila. Nature 285, 403-405.

LOHS-SCHARDIN, M., SANDER, K., CREMER, C , CREMER, T. AND

ZORN, C. (1979). Localized ultraviolet laser microbeamirradiation of early Drosophila embryos: Fate maps based onlocation and frequencies of adult defects. Devi Biol. 68,533-545.

MADHAVAN, M. AND SCHNEIDERMAN, H. A. (1978). Histologicalanalysis of the dynamics of growth of imaginal discs andhistoblasts nests during the larval development of Drosophilamelanogaster. Wilhem Roux Arch, devl Biol. 183, 269-306.

MORATA, G. AND KERRIDGE, S (1981). Sequential functions of thebithorax complex of Drosophila. Nature 290, 778-781.

NOTHIGER, R. (1972). The larval development of imaginal discs. InThe Biology of Imaginal Discs (ed. H. Ursprung and R.Nothiger) pp. 1-34. Berlin: Springer Verlag.

PATEL, N., SNOW, P. AND GOODMAN, C. (1987). Characterization

and cloning of fasciclin III: a glycoprotein expressed on a subsetof neurons and axon pathways in Drosophila. Cell 48, 975-988.

POULSON, D. (1950). Histogenesis, organogenesis anddifferentiation in the embryo of Drosophila melanogaster.Meigen. In Biologv of Drosophila (ed. M. Demerec) pp.168-270. New York: Hafner.

SHEARN, A. AND GAREN, A. (1974). Genetic control of imaginaldisc development in Drosophila. Proc. natn Acad. Sci. U.S.A.71, 1303-1397.

SIMCOX, A., ROBERTS, I., HERSPERGER, E., GRIBBIN, C , SHEARN,

A. AND WHITTLE, R. (1989). Imaginal discs can be removedfrom cultured embryos mutant for the segment polarity genesengrailed, naked, and patched, but not from wingless.Development 107, 715-722.

TRUMAN, J. AND BATE, M. (1988). Spatial and temporal patterns ofneurogenesis in the central nervous system of Drosophilamelanogaster. Devi Biol. 125, 145-157.

WIESCHAUS, E. (1978). Cell lineage relationships in the Drosophilaembryo. In Genetic Mosaics and Cell Differentiation (ed. W.Gehring) pp. 97-118. Berlin: Springer Verlag.

WIESCHAUS, E. AND GEHRING, W. (1976a). Clonal analysis ofprimordial disc cells in the early embryo of Drosophilamelanogaster. Devi Biol 50, 249-263.

WIESCHAUS, E. AND GEHRING, W. (1976£>). Gynandromorphanalysis of the thoracic disc primordia in Drosophilamelanogaster. Wilhem Roux Arch, devl Biol. 180, 31-46.

WIESCHAUS, E. AND NOSSLEIN-VOLHARD, C. (1986). Looking atembryos. In Drosophila/ a Practical Approach (ed D. Roberts)pp. 199-227. Oxford: IRL Press.

{Accepted 27 March 1991)