Search

The title of this Essay embodies an anachronism, because the term aneuploidy was not coined until several years after Theodor Boveri’s death. But that term would be meaningless if all chromosomes were functionally equivalent, and consequently no such term existed at the time when Boveri demonstrated the so-called individuality of chromosomes for the first time, by analysing dispermy in sea urchins. It is his exploitation of this “experiment of Nature” and its developmental implications that we shall now discuss, returning thereby from the role of the egg cytoplasm in development (see the previous Essay) to the roles that the nucleus, and more specifically the chromosomes, play. The work reviewed in this Essay and the next was, according to E.B. Wilson (1918, p. 74/75), Boveri’s “crowning achievement, whether in respect to excellence of method or importance of result.”

Gradients, for decades a “dirty word” in developmental biology (Lawrence 1992), have become respectable of late and are likely to stay with us after their role in pattern formation has been demonstrated at the level of gene regulation (Driever and Nusslein-Volhard 1988). The present essay is to illuminate the often misunderstood “relativistic” gradient concepts of Theodor Boveri (who actually introduced the term Gefalle = gradient), and their influences on his students. Boveri also paved the way, by his studies on chromosomes and development (see Essays 15 and 16), for a chromosomal location of the Mendelian hereditary factors (since 1909 called genes). His most influential students reacted quite differently to his insights on both gradients and genes. Hans Spemann (1869–1941) was remarkably reluctant in linking either con-. cept with his organizing center (see Sander 1986; Wolpert 1986). Waldemar Schleip (1879–1948), however, extended Boveri’s idea of interplay between cytoplasmic factors and genes in development, yet maintained a specific “Boverian” definition of gradient. Leopold von Ubisch (1885–1965) finally achieved a synthesis linking the emerging modern concept of morphogenetic gradients with differential gene activity in order to generate spatial patterns.

To obtain a clearer understanding of the evolutionary transition between short- and long-germ modes of embryogenesis in insects, we studied the expression of two gap genes hunchback (hb) and Krüppel (Kr) as well as the pair-rule gene even-skipped (eve) in the dipteran Clogmia albipunctata (Nematocera, Psychodidae). This species has features of both short- and long-germ mode of embryogenesis. In Clogmia hb expression deviates from that known in Drosophila in two main respects: (1) it shows an extended dorsal domain that is linked to the large serosa anlage, and (2) it shows a terminal expression in the proctodeal region. These expression patterns are reminiscent of the hb expression pattern in the beetle Tribolium, which has a short germ mode of embryogenesis. Krüppel expression, on the other hand, was found to be rather similar to the Drosophila expression, both at early and late stages. eve expression starts with six stripes formed at blastoderm stage, while the seventh is only formed after the onset of gastrulation and germband extension. Surprisingly, no segmental secondary Eve stripes could be observed in Clogmia although such segmental stripes are known from higher dipterans, beetles and hymenopterans. We therefore also studied another nematoceran, Coboldia, to address this question and found that some segmental stripes form by intercalation as in Drosophila, although belatedly. Our results suggest that Clogmia embryogenesis, both with respect to morphological and molecular characteristics represents an intermediate between the long-germ mode known from higher dipterans such as Drosophila, and the short-germ mode found in more ancestral insects.

We have studied the binding pattern of antibody mAB 2B8 directed against even-skipped orthologous proteins (EVE) in honeybee embryos. Primary and secondary EVE stripes form in roughly anterior-to-posterior succession; there are 8 primary and 16 secondary stripes. The most posterior primary stripes appear only after the onset of gastrulation. The secondary stripes form by a splitting of primary stripes; they demarcate the parasegmental pattern. While these findings resemble EVE expression in long-germ beetles, the honeybee differs from both beetles and dipterans by two transient pair-rule traits in the parasegmental EVE pattern: the secondary stripes in head and thorax alternate in strength, yet out of register with the Drosophila pattern, and over the whole pattern the odd-numbered stripes vanish earlier than their even-numbered counterparts. As in Drosophila, however, the strong EVE stripes coincide with the weak engrailed (EN) stripes. These findings are taken to indicate that (1) honeybee and beetles share a conserved mode of EVE stripe formation whilst Drosophila has diverged in this respect, (2) honeybee and Drosophila have diverged from the beetles in specific pair-rule traits during the parasegmental expression of both EVE and EN, and (3) some of these traits differ in the register of segment pairing and thus may reflect regulatory divergences at the pair-rule level between dipterans and the honeybee.

This award lecture reviews concepts of embryonic pattern formation developed since the time when, early in this century, insect embryogenesis first became a topic of causal research. These concepts, considered here in their historical sequence, are: (1) preformed patterns of determinants in “mosaic eggs”; (2) “regulative” pattern formation directed by a differentiation centre; (3) pattern specification by antagonistic polar gradients; and (4) the complex interactions of developmental genes recently revealed by genetical and molecular methods. These methods have triggered an epoch-making breakthrough in developmental biology, but they also can reveal surprising evolutionary aspects.

Blocking epiboly during gastrulation in teleosts and amphibians leads to so-called ring-embryos that have a uniform head and tail end while the trunk consists of two separate lateral halves forming a ring around the persisting yolk plug. Comparable malformations probably due to delayed gastrulation can be traced throughout the chordates from amphioxus to man, revealing a fundamental unity in morphogenesis. The present contribution describes the discovery and early interpretation of ring-embryos in the Anamnia and recalls a mistaken fate map derived from them that must have substantially delayed the discovery of the amphibian organizer effect.

The engrailed expression in embryos of a beetle, four midges and a fly has been analysed with special reference to the terminal regions. In all six species the segmental expression pattern is very similar but variability occurs in the clypeolabrum, foregut and hindgut. In some cases, segmental engrailed expression seems to be extended into the hind- and/or foregut. The engrailed expression of these species is compared with published data from other insects.

We discovered extremely small genomes (1C similar to 100 Mb) in the dipteran insects Coboldia fuscipes (Scatopsidae) and Psychoda cinerea (Psychodidae). The small genomes of these species cannot be explained by a fast developmental rate, which has been shown to correlate with small genome sizes in animals and plants but might accommodate the combined effects of other developmental traits, including small egg size, thin blastoderm layer, and long-germ development.

We discuss the interplay between evolution and development as reflected in data and concepts since about 1800. Darwin and his “continental apostle” Haeckel put the striking similarity between early vertebrate embryos in an evolutionary context. Haeckel's partly illicit generalizations discredited evolutionary thinking among early experimental embryologists who moreover noted riddles incompatible with contemporary concepts of homology and evolution. Relevant solutions were suggested by the more recent concept of ontogenetic networks that embody complex regulatory properties and genes with partly overlapping functions. Molecular data on development increasingly reveal evolutionary opportunism, for instance when a widespread signaling chain involved in primitive immune defense was apparently recruited later on for dorso-ventral axis determination in some evolutionarily advanced insect groups. Recently, Rickettsia-related bacteria colonizing many arthropod species were found to “manipulate” the first steps of host development to the advantage of their own propagation, but by ways that could also promote host speciation. Molecular genetics can now document evolutionary steps in ontogenetic networks. In the fruit fly for instance, the novel bicoid gene has superseded a crucial patterning function within a pre-existing network—a case of “molecular caenogenesis.” The expression patterns of conserved genes that antagonistically determine dorso-ventral polarity support a literal revolution envisioned almost 200 years ago. This is the dorso-ventral inversion of the body plan in some metazoans—ascribed then to the Articulata, now to the Chordata. The final section posits that the opportunistic character of evolutionary innovations is detrimental to parsimony in development. J. Exp. Zool. (Mol. Dev. Evol.) 302B:69–91, 2004. © 2004 Wiley-Liss, Inc.

Ernst Haeckel (1834-1919) must count among the most widely known biologists of his time. His monographs on radiolarian skeletons, sponges and medusae immediately became standard works, owing partly to lavish illustrations that later on culminated in his "Art Forms in Nature", which markedly influenced the "art nouveau" of the early 20th century. Haeckel's main impact, however, came from his numerous popular books that were crucial in transferring Darwin's ideas to continental Europe. Haeckel's main addition was his claim that ontogeny recapitulates phylogeny. It was partly founded on pre-Darwinian observations by J.F. Meckel and K.E. v. Baer who noticed that vertebrate embryos of different species resemble each other more strongly during early ontogenesis than later on. Wishing to illustrate this clearly, Haeckel clandestinely generalized some figures showing early embryos of animals and Man. This "fraud" provided ammunition for his many adversaries, most of whom felt provoked by his antireligious campaigns. The resulting controversies continued well into the 20th century but then subsided. Quite recently, however, they have flared up again, perhaps in connection with progress in molecular embryology that revealed an amazing evolutionary conservation of genes and their cooperation in signal transduction chains. The scientific publications that triggered this flare, and a selection of "Letters to the Editor" in both international science magazines and the German popular press, serve here to show that Haeckel's idea of recapitulation, while having proven its heuristic value, is still causing considerable irritation. This results from the widespread intuition that the marvels of ontogenesis and other biological phenomena must reflect some "intelligent design" rather than Darwinian evolution.

As described in the previous Essay, Theodor Boveri, by exploiting the “natural experiment” of sea urchin dispermy, demonstrated that embryogenesis requires “a certain combination of chromosomes”, possibly the full chromosome complement. The main topic of this Essay is his reflections on the methods he employed, and on their value for gaining previously inaccessible insights.

Theodor Boveri, in the words of Edmund B. Wilson (1918), “throughout his life held with extraordinary tenacity to the investigation of a group of problems that were already clearly in his view during his first researches. What is the egg, and what does it contain? How are hereditary traits represented in the egg and determined during its development? What is the background from which they emerge into view, one after another, as development goes forward? By what mechanism (if mechanism it be) do the unseen potentialities hidden within the germ become realities that appear before our eyes in the embryo and adult?”

To realize his plans to eliminate individual blastomeres of the ascidian embryo (see Fischer 1992a, b), around 1885 Laurent Chabry devised a sophisticated set of micro-tools for use under the microscope. It is the aim of this essay to describe these tools, to explain his methods of constructing and employing them and to follow their possible influences on subsequent research and on tools developed by later experimental biologists.

In the very year, 1883, when Wilhelm Roux was pondering on the meaning of mitosis (see Sander 1991), he started experimenting on amphibian eggs. The first problem he set out to solve — or rather, the first question he discussed in writing — was the orientation of the prospective embryo in the egg. In amphibians this is a formidable problem, partly owing to the highly complex gastrulation movements, which in some respects have remained enigmatic to this day. Indeed, on trying to ascertain dorsoventral and antero-posterior polarity in the frog’s egg, Roux was confused by aberrant gastrulation movements in quite a spectacular way. Irrespective of this, he gained remarkable and still valid insights concerning the prospective median plane in frog eggs and the epigenetic factors that bring about its orientation — the latter study (Roux 1887) being adorned with further speculations on equal and unequal mitoses during early cleavage.

A number of genes of the developmental gene hierarchy in Drosophila encode transcription factors containing Cys2His2 zinc finger domains as DNA-binding motifs. To learn more about the evolution of these genes, it is necessary to clone the homologs, or more correctly the orthologs, from different species. Using PCR, we were able to obtain apparently orthologous fragments of hunchback (hb), Krüppel (Kr), and snail (sna) from a variety of arthropods and partly also from other animal phyla. Sequence alignments of these fragments show that the amino acid differences can normally not be correlated with the evolutionary distances of the respective species. This is due to an apparent saturation of potential replacements within the finger domains, which is also evident from the frequent occurrence of convergent replacements. Another recurrent feature of these alignments is that those amino acids that are directly involved in determining the DNA-binding specificity of the fingers are most conserved. Using in vitro bandshift experiments we can indeed show that the binding specificity of a hunchback finger fragment from different species is not changed. This implies that there is a high selective pressure to maintain the regulatory target elements of these genes during evolution.

The transition from oogenesis to embryogenesis has received little attention in insects so far, owing largely to internal fertilization that prevents direct observation. Fertilization is characterized by extreme sperm economy, sometimes linked to gigantic sperm length. Egg activation as a rule is not initiated by sperm entrance but by changes connected with egg deposition (deformation, water uptake etc.). Both fertilization and activation can be achieved in vitro. The ooplasmic movements prominent in many species during both oogenesis and early embryogenesis seem largely independent of cleavage nuclei but may require extranuclear sperm contributions.

In the realm of biology, Hans Driesch (1867–1941) is known for his early experiments on sea urchin embryos (see the previous Essay), and for his later credo that life —and particularly ontogenesis — depends on some immaterial vital force and thus cannot be fully understood in terms of physics and chemistry; that indeed it defies any materialist approach (the topic of the next Essay). What is rarely mentioned are his early theoretical writings on embryogenesis, only traces of which — such catchwords as prospective value and prospective potency,and the Fundamentalsatz that a cell’s fate is a function of its position — have survived, the last owing to its happy re-incarnation in the mind of Lewis Wolpert (Wolpert 1986).

Among the early adherents of Entwickelungsmechanik Hans Driesch (1867–1941) was certainly the most colourful personality. Born at Kreuznach, a well-known spa on the Nahe river, as the only child of a wealthy Hamburg merchant, he grew up in the liberal atmosphere of that hanseatic city (Driesch 1951). Gifted with the self-confidence that results from a strong mind backed by seemingly unlimited financial means, and driven by a brillant intellect formally trained at the best institutions within reach, he had by his mid-twenties published nearly two dozen descriptive and experimental papers on animal development, and three small books in the border-fields between theoretical biology and philosophy of science. Many of these have fallen justly (and some unjustly) into oblivion, while two or three proved seminal. These last dealt with the shaking and compressing of cleavage stage sea urchin eggs, and served more than any others to shake confidence in the then widely accepted concept of mosaic development (see the previous essays in this series).

Our SEM study of honeybee, Apis mellifera (Hymenoptera : Apidae), embryogenesis is based on embryos fixed at 1 hr intervals from oviposition to hatching. Embryos of equal age showed little variation, so that staging could be based on developmental age. Our data confirm many earlier light microscopical observations, but are at variance with some others. The cytoplasmic connections between the future blastoderm cells and the central yolk system are severed only at the onset of gastrulation. The serosa derives from cells which immigrate into the dorsal strip and then join up to form a pre-serosa bordering the germ band rims. When the serosa has detached, the amnion grows out from the germ band margins and serves as a provisional dorsal epithelium right from the beginning. Germ band segmentation is followed by the transient regression of every second transverse groove (double segment pattern). The germ band flanks grow dorsally and replace the amnion a few hr before hatching (dorsal closure). The tracheal openings which form half-way between segment borders are closed temporarily by the embryonic cuticle; similar openings above the labial buds contribute to the tentorium rather than the tracheal system. Most head appendages retain bud character until long after hatching. The events observed in the SEM are linked in a diagram to the stage series based on living embryos.

The aberrant pattern, “double abdomen,” previously induced in the egg of Smittia by uv irradiation of anterior pole regions was also produced by puncturing of the egg at the anterior pole. Double abdomens and embryos with anterior defects developed in eggs in which puncturing had locally prevented the regular arrangement of cleavage nuclei in the periplasm. The resulting gap in the blastoderm at the anterior pole was subsequently closed under exclusion of a small amount of egg material. Double abdomens did not develop in eggs where exclusion of anterior egg material was not observed. Thus a basic switch in the developmental program of the egg appears to depend upon the functional elimination of some crucial components in the anterior egg region.

We measured the intracellular electrical potential in oocyte and nurse cells of Drosophila follicles at different developmental stages (6–14) and determined the intrafollicular potential difference. During stages 8–10B, when intrafollicular transport is known to occur, no significant potential difference was found. During late vitellogenic stages the nurse cells assume a more positive potential than the oocyte. This result contrasts with the published data on Hyalophora follicles, in which intercellular electrophoresis of negatively charged proteins occurs from nurse cells to oocyte as a result of an intrafollicular potential difference (nurse cells more negative than the oocyte). Such a potential difference was not observed in Drosophila follicles at any stage, not even after application of juvenile hormone. The extrafollicular electrical field is described with a dipole model. The hypothetical dipole is located in the long axis of the follicle and changes its calculated length stage-specifically.

Summary Mature chorionated oocytes (eggs) dissected from ovaries of the dipterans Psychoda and Smittia can be activated by brief exposure to de-ionized water, tap water or low osmolality saline. Under optimal conditions larvae will hatch from nearly 100% of the experimentally activated eggs. In Psychoda, the yield of activation decreases with increasing osmolality of the activating medium; the threshold is near 350 mosmol/1, well below the osmolalities of Grace's and Robb's tissue culture media. The temporal threshold for activation is above 3 min and below 5 min. The effects of water treatments lasting up to 1 min are completely reversible; treatments of 2 or 3 min caused irreversible damage while failing to initiate embryogenesis. Storage of Psychoda eggs in Grace's medium for up to (at least) 6 h has no adverse effects on subsequent experimental activation while storage for 24 h reduces the fraction of activable eggs and also tends to impair embryogenesis. The data presented here indicate that low osmo...

Summary Eggs taken from the genital tracts of Venturia (Nemeritis) canescens (Hym., Ichneum.) and Psychoda cinerea (Dipt., Psychod.) can be activated by brief exposure to distilled water. However, embryogenesis fails soon unless some other conditions are met. These are mechanical deformation of the egg in Venturia, and mating in Psychoda. Development ceases at or soon after meiosis in Psychoda and in most Venturia eggs; the remaining Venturia eggs fail to form an orderly blastoderm although they contain hundreds of nuclei. In Venturia the anomalies are probably due to insufficient activation of some cytoplasmic component(s) while Psychoda eggs fail in the absence of some paternal contribution, possibly the sperm which in this case must enter the oocyte within the follicle.

Drosophila females homozygous for the mutation dicephalic occasionally produce ovarian follicles with a nurse-cell cluster on each oocyte pole (dic follicles). Most dic follicles contain 15 nurse cells as in the normal follicle, but the total nurse-cell volume is larger in dic follicles; this is in keeping with the increase in DNA content recently described. However, the relative increase in oocyte volume during nurse-cell regression (from stage 10B onward) is not significantly larger in dic than in normal follicles. Time-lapse recordings in vitro show that, as a rule, both nurse cell clusters in a dic follicle export cytoplasm to the oocyte but nurse-cell regression remains incomplete at both poles and the persisting remnants of the nurse cells cause anomalies in chorion shape. The kinematics of cytoplasmic transfer are less aberrant at that oocyte pole which harbours the germinal vesicle. Possible links are discussed between these anomalies of oogenesis and the double-anterior embryonic patterns observed in the majority of developing dic eggs.

Medial clefts in the axis of the trunk region are malformations known from many chordates and are mostly referred to as rachischisis anterior. In teleosts, rachischisis was previously ascribed either to secondary rifting of a single uniform axial rudiment, or to the establishment of two (half) axes and body halves physically separate from the very beginning. In order to decide between these conflicting interpretations, we treated zebrafish embryos during blastodisc stages and epiboly with several chemical teratogens causing rachischisis anterior. Treatment with ethanol, Colcemid, hydroxyurea, or cycloheximide was found to delay the proliferation and movements of the deep cells more strongly than the timing of cell differentiation, so that the deep cells embark on organogenesis before having reached their destinations in the uniform germ shield. Treatment with alpha-amanitin, on the other hand, seems primarily to affect the periblast and enveloping layer; the incomplete epiboly observed in these layers appears to restrain deep cell epiboly physically and thus to cause rachischisis. In both instances, the split condition of the embryo's trunk region is clearly due to the ectopic formation of physically separate body halves right from the beginning, a mode we call bipartite axiation. We also describe secondary anomalies specific for individual teratogens, and briefly discuss the possible origins of rachischisis anterior among other chordates including man.

Drosophila embryos were locally irradiated with a 257-nm laser microbeam during blastoderm and germ band stages. Depending on stage and beam diameter (10–30 μm), from 0 to 45 nuclei were exposed to the uv radiation. The doses used, 5 or 10 erg, did not eliminate nuclei or cells at once, but up to 50% of the adult survivors from irradiated eggs carried defects in the thorax. These were scored with reference to the imaginal discs from which the affected structures derive. For each thoracic disc a “target center” was calculated as the weighted mean value of all beam locations affecting the respective adult derivatives. The target centers for the germ band stage map within the respective germ band segments. The pattern of target centers for the blastoderm stage is comparable to the thoracic region of published fate maps, and the distances between adjacent leg centers (approximately three cell diameters) agree with recent evidence based on mosaic flies. We discuss the question whether the target centers mark the position of the respective disc progenitor cells at the stages of irradiation and conclude that these positions are rendered rather correctly at least with reference to the longitudinal egg axis.

U.v. irradiation of the anterior pole of nuclear multiplication stage Drosophila eggs produces embryos with defective anterior structures. At a low frequency embryos resembling some phenotypes of the bicaudal syndrome of Drosophila were observed. These embryos had no head or thorax and the eight abdominal segments were spread to the anterior of the embryo. Sometimes spiracles, characteristic of the most posterior embryonic segment were observed at the anterior of the embryo. The development of these embryos was followed, and abnormalities occurred as early as blastoderm formation. The extent of the blastoderm defects correlated well with the final abnormality in the embryo.

In the Drosophila ovarian follicle, 15 polyploid nurse cells (NC) are linked by cytoplasmic bridges to the anterior pole of the oocyte. The mutant dicephalic occasionally produces aberrant follicles whose 15 nurse cells are separated into two clusters, each attached to one oocyte pole [split nurse cell (SNC) follicles]. Feulgen cytophotometry of isolated nuclei from such clusters and from normal follicles of the same females revealed an inverse correlation between NC number per cluster and mean DNA content per nucleus. The maximum DNA content of nuclei from split NC clusters falls into the 4,096 C class as opposed to the 2,048 C class in 15 NC clusters; the mean DNA content per NC nucleus is 1,940 C in the former and 1,340 C in the latter. These correlations are confirmed by measurements on follicles containing fewer or more than 15 NC, and on SNC follicles from another mutant (spindel C). These findings indicate that replication in NC nuclei is controlled with reference to the transcriptional capacity of the individual NC cluster rather than the total capacity available per oocyte or follicle. We discuss this finding with respect to embryonic patterning as influenced by nurse cell location and list and discuss conflicting published reports on over- and underreplication in Drosophila follicles.

We determined the extracellular electrical current pattern around Drosophila follicles at different developmental stages (7–14) with a vibrating probe. At most stages a characteristic pattern can be recognized: current leaves near the oocyte end of the follicle and enters at the nurse cells. Only at late vitellogenic stages was an inward-directed current located at the posterior pole of many follicles. Most striking was the observed heterogeneity both in current pattern and in current density between follicles of the same stage. Different media (changed osmolarity or pH, addition of cytoskeletal inhibitors or juvenile hormone) were tested for their effects on extrafollicular currents. The current density was consistently influenced by the osmolarity of the medium but not by the other parameters tested. Denuded nurse cells (follicular epithelium locally stripped off) show current influx, while an accidentally denuded oocyte produced no current. Our results show that individual follicles may be electrophysiologically different, though their uniform differentiation during vitellogenesis does not reflect such heterogeneity.

The prosobranch snail Assiminea grayana is living on muddy shores at a level flooded only during spring tides. Its spawn consists of 2-layered capsules containing one egg cell each. They are deposited in clusters protected by a cap formed of faecal pellets. The veliger larva ready to hatch may survive within the egg capsule for several months until suitable hatching conditions occur. These are provided by lengthy or

Some details concerning the reproduction of 2 Indian species of the supralitoral prosobranch genus Assiminea are described and compared with those of a North Sea species. The eggs are deposited on the mud surface and covered by an oblong cap formed of faecal pellets. The transversal dimension of the cap is correlated with the species-specific size of the pellets, the length corresponds to the number of eggs covered. In the North Sea Assiminea the cap is more or less straight, in the Indian species it meanders or is coiled up. The free-living veliger stage in both tropical species is more extended than in the North Sea species; this observation agrees with the rule of Thorson.

UV-irradiation (maximum intensity at 253,7 nm) of the anterior 1/3 of the egg during nuclear migration stage frequently is followed by the formation of "double abdomens". The type of malformation shows a germ band in which head, thorax and anterior abdominal segments are replaced by a set of 6-8 posterior abdominal segments joined in mirror-image symmetry to the corresponding parts of the original abdomen. The course of development leading to this malformation is visibly aberrant from the blastoderm stage to germ band segmentation only. The two abdomens develop in strict synchrony provided their positions are symmetrical to the equatorial plane of the egg. Asymmetric positions are less frequent and delay temporarily the development of one or the other of the abdomens due to the lack of space. The results are compared to similar malformations in other insect eggs and possible interpretations are reviewed.

Insect segments are seriated developmental subunits whose borders largely control intra-segmental patterning. The individual segment owes its specific character to position-dependent instructions given to its early embryonic progenitor cells. In Drosophila, the segment borders first become visible as a series of transverse infoldings on the surface of the germ band. We describe here development in a mutant (prdFR1) which forms half the normal number of transverse infoldings spaced at twice the normal distance, thereby subdividing the germ band into a series of giant domains (Fig. 1 a-c). Each domain develops the tracheal pits of two normal segments (Fig. 1e), thus revealing its composite origin, but the larva expresses only one segment border per domain and the cuticle between these borders is similar to the normal segment in both size and zonation (Fig. 2). We conclude that this course of development reflects a conflict between the seriated instructions specifying local segment character (see Fig. 1b) and controlling influences exerted by the abnormally spaced infoldings and later by the definitive segment borders.

Publisher Summary This chapter focuses on the problem of pattern specification in early insect embryogenesis, essentially on the formal level characteristic of “classical” developmental physiology. In the field of pattern formation, optimal use can be made of these new methods of study only if they are applied with the proper questions in mind; starting off from the wrong hypothesis may mean years of detour. This consideration requires integrating old and new data from an animal group, which, such as the insects, holds much promise for the developmental biologist, especially from the genetical point of view. Any embryonic cell or group of cells, which later on, will give rise to a particular part of the body, must be instructed by appropriate signals, which specific pathway of development they should follow; this process will be called cellular instruction. Specification of a pattern occurs when different instructing signals are provided in a spatially coordinated manner. In order to respond to a specific instructing signal by producing a particular visible pattern element, an embryonic cell or group of cells must become committed to the developmental pathway indicated by that signal. Much evidence concerning pattern specification is derived from the study of egg fragments; parts of the system isolated functionally by removal or denaturation (cautery) of the remaining parts.

A monster embryo with a head on each end developed from an abnormalDrosophila egg produced by a M(1)oSp/Binsn female. The embryo probably represents a longitudinal mirror duplication of head and all 3 thoracic segments. The significance of this type of pattern anomaly for embryonic pattern specification inDrosophila is discussed.

The pattern of intercellular connections between germ line cells has been studied in follicles of the mutantdicephalic (dic), which possess nurse cell clusters at both poles. Staining of follicles with a fluorescent rhodamine conjugate of phalloidin reveals ring canals and cell membranes and thus allows us to reconstruct the spatial organization of the follicle. Each germ line cell can be identified by the pattern of cell-cell connections which reflect the mitotic history of individual cells in the 16-cell cluster. The results indicate that in both wild-type anddicephalic cystocyte clusters one of the two cells with four ring canals normally becomes the pro-oocyte. However, in some follicles (dicephalic and wild-type) oocytes were found with fewer or more than four ring canals. Indic follicles, one or several nurse cells may become disconnected from the other cells during oocyte growth at stage 9-10. Such disconnected cells cannot later on empty their cytoplasm into the oocyte. This, in turn, might be of consequence for the determination of axial polarity of the embryo.

The segmental pattern of insect embryos depends on influences from morphogenetic centres near each of the egg poles1. In Drosophila, maternal effect mutations are known that impair the normal function of each centre2. Injection of wild-type cytoplasm into mutant eggs has revealed that morphogenetic signals localized at the anterior and posterior pole of eggs can be transplanted3,4. We show here that these activities can also be detected during oogenesis. Posterior activity can be recovered at an early stage (stage 10, ref. 5) from the oocyte-nurse cell complex, but anterior activity can only be detected in the mature oocytes (stage 14). We conclude that the bicoid-dependent anterior signal, although produced by the nurse cells, does not become active before it is localized to the anterior egg pole, whereas posterior activity can be detected in the nurse cells before, and therefore independently of, its localization to the posterior egg pole.

Investigations on pterygote insect oogenesis have yielded a multitude of data on the architecture and development of early and late oocytes (for reviews see King, R. C., 1970; Telfer and Smith, 1970; Mahowald, 1972; Anderson, 1974; Telfer, 1975; see also Miya et al., 1969; Truckenbrodt, 1970; Buning, 1972). Oogenesis has also been studied in some Apterygota (Cone and Scalzi, 1967; Cantacuzene and Martoja, 1972; Palevody, 1972, 1973; Matsuzaki, 1973; Bilinsky, 1976, 1977, 1979; Klag, 1977, 1978). Yet the end product of oogenesis was usually exempted from investigation; the newly laid insect egg thus remains largely unknown with respect to ultrastructure. This may mainly be due to technical reasons: the prevalence of yolk and the extremely resistant egg covers of insect eggs provide considerable handicaps for fixation and sectioning. Descriptions of the egg shell are therefore much more frequent than publications dealing with the egg cell.

We have isolated and characterized a homeoboxcontaining gene from the honeybee Apis mellifera. Its homeobox region shows a high degree of sequence similarity to the homeobox of the Drosophila gene Deformed (Dfd). At the DNA level 82% of the basepairs are the same, whereas the putative amino acid sequences are identical between the bee and the fruitfly genes. Similarity is also present 5′ and 3′ to the homeobox. Using this isolate as a probe we have performed in situ hybridization on sections from blastoderm-stage embryos of the honeybee Apis mellifera. In early blastoderm stages we found a rather irregular pattern of labelled nuclei. In middle stages we found silver grains over each nucleus and also over the cytoplasm in a belt of blastoderm cells in the prospective gnathal region. These results indicate that the Deformed genes from honeybee and fruitfly are homologous both with respect to their DNA sequence and their spatial and temporal pattern of expression during embryogenesis.

Mutations altering the spatial pattern of embryonic primordia are valuable tools for studying the earliest steps of embryonic pattern formation. As the spatial organisation of the egg cell arises during oogenesis, such mutants are likely to affect the embryo via the maternal genome (maternal effect). In Drosophila, most maternal-effect mutants studied so far lead to early developmental arrest. A few others yield pattern defects apparently linked to local failure of blastodermal cellularisation rather than to abnormal arrangement of primordia or ooplasmic determinants1,2. A small number of mutants, however, are known in which the embryonic fate map is affected, for example, bicaudal3,4, which profoundly alters the antero-posterior pattern of embryonic organisation. The bicaudal phenotypes suggest that the mutation causes a perturbation of a morphogen gradient that defines the antero-posterior coordinate in early embryonic pattern formation5,6. A new maternal-effect mutant, dorsal, seems to exert a similar influence on the dorso-ventral coordinate for patterning6. Dorsal (dl) has both a recessive and a dominant maternal-effect phenotype. The dl recessive phenotype is marked by an apparently complete lack of dorso-ventral polarity from early embryogenesis onward6. The dl dominant (dlD) phenotype, which we report on here, is temperature sensitive: at 22 °C, dl/+females lay eggs which develop normally while eggs produced at high temperature (29 °C) yield abnormal larvae which fail to hatch. These mutant dlD larvae vary in phenotype but nearly always lack muscles and frequently part of the ventral hypoderm. They are defective in structures that arise from the ventral region of the blastoderm in normal embryogenesis6,7 (Fig. 1).

1. Eggs of the leaf hopperEuscelis plebejus F. were fragmented longitudinally by a novel technique during cleavage. 2. Irrespective of the plane of fragmentation, both fragments may produce germ bands which are either of perfect bilateral symmetry or contain at least some bilaterally symmetrical body segments. With one possible exception, no longitudinal half-patterns were formed. 3. Ventral and lateral egg halves produce either complete germ bands or anterior partial germ bands. The failure to produce posterior body segments is correlated to the absence of special posterior pole materials marked by the symbiont mass. 4. Dorsal egg halves may produce either complete germ bands or posterior partial germ bands. The loss of anterior segments may be due to lack of keimanlagen cells. 5. The results are discussed and compared to those from transverse fragmentations carried out earlier. 6. Aided by a formalized presentation of the results a dilemma of definition is pointed out which results from the common usage of defining embryonic regulation as the alternative of mosaic development and at the same time by reference to the normal result of development.