Following cauterization in different regions of each larval segment, different effects - on growth and pattern - in the adult hemitergite were obtained. A correspondence between different segments in the larva and in the adult was found, with the exception of the Vllth abdominal segment in the male, which apparently has no adult cuticular representation. The histoblasts forming the imaginal Anlage are grouped in the central part of the larval hemisegment. Following cauterization, the viable histoblasts only partially replace the killed ones, giving rise to incomplete tergites.

Pattern damages are characteristic of the cauterized region. Cauterization in marginal regions of the abdominal segment, although not killing the histoblasts, induce specific fusions and penetrations between neighbouring colateral and contralateral hemitergites. The effects of cauterization on the number of cuticular elements and their disposition in the adult pattern are discussed.

Clonal analysis of the different imaginal discs of Drosophila has recently provided a closer insight into the developmental parameters of these embryonic systems. Among the studied imaginal discs, the Anlagen giving rise to the adult tergites show various exceptional features. Whereas the cephalic (eye, Becker, 1957; antennal, Postlethwait & Schneiderman, 1971) and thoracic discs (wing disc, García-Bellido, 1968; Bryant, 1970; García-Bellido & Merriam, 1971 a; legs, Bryant & Schneiderman, 1969) grow by cell proliferation during the larval development and cease to grow after pupariation, the tergite Anlagen (Lobbecke, 1958, García-Bellido & Merriam, 1971 b) remain quiescent throughout the larval development, starting proliferation after pupariation. Moreover, while the cephalic and thoracic discs develop in isolation from the remaining larval epidermis, the histoblasts of the tergite Anlagen proliferate and expand among the larval epidermal cells to fill the presumptive tergite area (Robertson, 1936). Finally, the adult pattern of cuticular elements in the tergites is more repetitive and simple than the complicated cuticular patterns of other imaginal discs.

The aim of the present work is (1) to locate experimentally the small clusters of histoblasts in each hemitergite, (2) to study some of the morphogenetic mechanisms leading to the proliferative growth and spreading of the histoblasts, and (3) to follow their subsequent organization into adult cuticular patterns. Our results will be compared with previous ones gained from Drosophila (Maas, 1948; Sobels, 1952; Zimmermann, 1954; Löbbecke, 1958) as well as from other insects (see revue in Lawrence, 1970).

Drosophila melangaster (wild-type Sevelen) individuals of different ages were used throughout the present experiments. The age of the larvae was calculated in hours (± 4) after egg-laying. The larvae were separated according to sex in experiments on the Vlth, VIIIth and Vlllth abdominal segments.

Cauterization was carried out with a tungsten thermocauter. The tungsten tip, tapered to 0-01 mm in diameter, was heated by the contact - 4 mm proximal to the tip - of two tungsten wires connected to both poles of a 1·5 V battery. In order to standardize cauterization, the heat developed by the tungsten tip was measured on paraffin (56°C melting point). By lightly applying the tungsten tip to the surface of the paraffin, the spot diameter of the melted paraffin increases with increasing heating time, reaching a plateau after 6 sec. The same diameter is reached by putting a preheated tip on contact with the paraffin for only 3 sec. The diameter of the melted paraffin, about 1 mm, corresponds to several larval segments. However, the heat transmitted to the epidermis is presumably lower than measured on paraffin, due to the lower conductivity of the cuticle. Cauterized larvae show after puparium formation damaged pupal cases in spots not exceeding the surface of a hemisegment.

Etherized larvae were placed on adhesive tape and cauterized in the desired larval segment with the preheated tip for 3 sec. A constant fraction of the treated larvae metamorphose and hatch as adults. Their abdomens were dissected out, cooked in 10% KOH, washed and mounted in Euparal for microscopic examination.

Gynandromorphs were obtained by the spontaneous loss of the Catcheside ring in females In(1)Xc2wvc/yf36 (see García-Bellido & Merriam, 1971 b). Mosaic spots were induced by 100 r X-rays (Philips 151 /Be, 150 rev/min, 100 kV, 15 mA, 2 mm A1 filter) in larvae of the genetic constitution y/ +; mwh/ + . Sb63/ + (see García-Bellido, 1972).

I Morphology of the larval and imaginal abdominal segments

The larva of D. melanogaster has one cephalic, three thoracic and eight abdominal segments. The different abdominal segments are externally marked by the alternation of bands of thin cuticle and bands of hard cuticle, showing circular rows of small cuticular processes or ‘hooks’ (Figs, 1a, 2d). The limits of these segments correspond to the internal attachment of the longitudinal muscles to the band of hooks. For the purpose of cauterization we subdivided the dorsal half of each larval hemisegment into 12 regions (Fig. 2d). Regions A, D, G and J are the most dorsal or medial, and C, F, I and L the most ventral or lateral. Regions C, B and A, the most cephalad of the segment, lie on the most anterior portion of the thin cuticle band. The regions J, K and L, the most caudal, lie on the band of hard cuticle with hooks.

Fig. 1

External morphology of the larva and the adult abdomen showing the correspondence between segments, (a) Larva, (b1) female, (b2) male abdomen. 1–3, Thoracic segments. I-VIII, abdominal segments.

Fig. 1

External morphology of the larva and the adult abdomen showing the correspondence between segments, (a) Larva, (b1) female, (b2) male abdomen. 1–3, Thoracic segments. I-VIII, abdominal segments.

The adult fly abdomen has seven segments covered dorsally by tergites (Figs. 1b, 2b). The eighth larval segment has no adult structure corresponding with tergites. Its derivatives are homologous with the genital arch, which originate from the genital imaginal disc (Ferris, 1950). Males differ from females in not having a separated seventh tergite. The sixth tergite of the male is larger and has two lateral spiracles, instead of one. This has lead to consider it as resulting from the fusion of the imaginal derivatives of both sixth and seventh larval segments (Dobzhansky, 1931; Stern, 1966).

The adult tergites show a highly constant pattern in the distribution of pigment and chaetes. Slight variations in this pattern are characteristic for each segment. The pattern of the Vth tergite is representative of tergites II to VI in both males and females (Fig. 2b). Between two neighbouring segments there is a thin and unpigmented band of cuticle corresponding to the intertergital fold. The rest of the tergite consists of a band of hard cuticle, uniformly covered by about 17000 trichomes, with yellowish pigment occupying the two-thirds most cephalad region and with a black pigment band covering the one-third most caudal region. Chaetes are present in the two-thirds caudal band of the tergite. These are of two classes, microchaetes (47·2, S.E. ± 7·3 per hemitergite, pooling the data of both males and females of the present stock), which appear regularly distributed over the entire surface, and macrochaetes (7·7 ± 1·4) which are restricted to a single row on the most caudal margin of the heavily pigmented band. The adult hemitergite was subdivided for analysis into nine regions as shown in Fig. 2b.

Fig. 2

Larval and adult regions in which each hemisegment was subdivided. (a) Larval cuticle, (b) adult tergite. ○ Microchaetes; ⊚, macrochaetes; V, hooks.

Fig. 2

Larval and adult regions in which each hemisegment was subdivided. (a) Larval cuticle, (b) adult tergite. ○ Microchaetes; ⊚, macrochaetes; V, hooks.

II The organization of the imaginal Anlage

(a) Localization of the imaginal histoblasts

We studied first the effects of cauterization of the left dorsal Vth abdominal segment. Cauterization in any of the 12 larval regions in which it was sub-divided (Fig. 2a) causes, in a fraction of the cases, larval or pupal mortality. However, this mortality seems not to be correlated with cauterization in a specific larval region (Table 1). The emerged adults show two main types of damage. Sometimes the adult left Vth hemitergite may show growth damage, lacking parts of the tergite or being completely absent. In other cases the partial or complete hemitergites may show pattern damage - that is, morphogenetic abnormalities in the number or disposition of the adult cuticular elements. In a fraction of the cases this damage may extend to colateral or contralateral neigh-bouring hemitergites. As shown in Table 1, the parameter c/b-that is, the fraction of emerged flies with defective hemitergites, is characteristic of the cauterized larval region. An analysis of the fraction of cases with total absence of the Vth left hemitergite, with remnants smaller than 50% or with lower degrees of reduction, shows that they are correlated with and typical of each region cauterized. That means that the existence of remnants and the total absence of the adult hemitergite both represent different degrees of damage in the imaginal anlage. The maximal growth damage is obtained following cauterization in the region F, close to B and E (Fig. 3). Thus, the histoblasts of the adult hemitergite may possibly be located into the upper right corner of the region F.

Table 1

Effects of cauterization in different regions of the Vth larval hemisegment

Effects of cauterization in different regions of the Vth larval hemisegment
Effects of cauterization in different regions of the Vth larval hemisegment
Fig. 3

Location and proliferative growth of the histoblasts. (a) Degrees of growth damage following cauterization in the larval segment regions (see Table 2). (b) Degrees as percentage in which the adult regions are populated in tergite remnants. Lines group regions of similar responsiveness.

Fig. 3

Location and proliferative growth of the histoblasts. (a) Degrees of growth damage following cauterization in the larval segment regions (see Table 2). (b) Degrees as percentage in which the adult regions are populated in tergite remnants. Lines group regions of similar responsiveness.

We tried to ascertain how closely packed are the histoblasts in the Anlage, assuming that the more clustered they are the more they will respond to cauterization as a unit. We classified the defective hemitergites in classes of 0 to 9/9 according to the fraction of remnant hemitergite surface and calculated the percentage of spots of these sizes which appear following cauterization of the different 12 larval regions. All the regions behave as a unit in so far that the majority of the resulting adult hemitergites correspond to the 0 or 9 classes (Fig. 4). Other size classes appear with very low and uniform frequencies on all the treated regions. Thus, the histoblasts respond mainly with an all-or-none reaction, indicating that they are clustered probably in a single nest and not scattered over a broad region.

Table 2

Effects of cauterization in regions E or F of different larval hemisegments

Effects of cauterization in regions E or F of different larval hemisegments
Effects of cauterization in regions E or F of different larval hemisegments
Fig. 4

Distribution of tergite sizes following cauterization in the different segment regions. Ordinate: percentage of the total number of tergites. Abscissa: tergite sizes in fractions of 0–9/9.

Fig. 4

Distribution of tergite sizes following cauterization in the different segment regions. Ordinate: percentage of the total number of tergites. Abscissa: tergite sizes in fractions of 0–9/9.

The existence of intermediate classes, fractions 1–8/9, indicates that regeneration during the population of the hemitergite is low. Otherwise we should have found only completely lacking or entire hemitergites. The degree of regeneration in the cauterized histoblast nest can be evaluated comparing the size of X-ray- induced clones in cauterized and non-cauterized hemitergites. Larvae of the genetic constituion y/ +; mwh/ + . Sb63/ + were irradiated, and once they reached the late third instar, cauterized in regions E and F of the Vth left hemisegment. In the emerged adults we analysed the number of y or Sb+ chaetes in spots of the Vth left cauterized hemitergite and in the control hemitergite III. Only spots with more than one chaete are considered. Whereas the fraction of the adult tergite occupied by the mosaic spot is 8·0% in the control hemitergite, this is 9·6% in the almost complete Vth hemitergite and 22·9% in defective Vth (less than 30 chaetes) hemitergites. This indicates that the size reduction of the cauterized tergite is due to killing of the histoblasts rather than to prevention of growth in the damaged Anlage. Moreover, whereas the mean number of chaetes for clones in the hemitergite III is 4·2 (n = 65) and 4·9 (n = 31) in Vth hemitergite with more than 30 chaetes, this is 4·4 (n = 8) in Vth hemitergites with less than 30 chaetes. These results suggest that possibly there is no regeneration or regulative growth in the cauterized Anlagen.

(b) Relationship between the larval and the adult segment

Cauterization in a given segment of the larva may lead to damage in the corresponding adult segment as well as in neighbouring segments. In order to find out whether a correspondence between the limits of the larval and the adult segment exists, we studied the effects of cauterizing in different regions of the Vth larval segment, along the cephalocaudal axis, upon the adult tergite of the IVth, Vth and Vlth segments, as well as the effects of cauterization of the IVth and Vlth segments upon the Vth tergite. As seen in Fig. 5, the proportion of defective tergites in the Vth tergite is lower, following cauterization in the hook region preceding the Vth segment, than after cauterization of the hook region in the posterior part of the Vth segment. Thus we conclude that each adult segment corresponds to the larval segment beginning after the cephalad band of hooks.

Fig. 5

Distribution of the frequency of growth damage following cauterization along the cephalocaudal axis on adjacent tergites. Ordinate: growth damage. Abscissa: effect of cauterization on different regions of each segment. □– –□, segm. IV; ○ —○, segm. V; △ – – △ segm. VI.

Fig. 5

Distribution of the frequency of growth damage following cauterization along the cephalocaudal axis on adjacent tergites. Ordinate: growth damage. Abscissa: effect of cauterization on different regions of each segment. □– –□, segm. IV; ○ —○, segm. V; △ – – △ segm. VI.

As seen before, cauterization on the regions E and F of the Vth larval hemi-segment leads to the highest proportion of defective tergites. In order to detect a correspondence between the larval segments and the adult tergites we cauterized the regions E and F of different segments and looked for specific effects in the adult tergites. These effects are classified according to larval mortality, growth damage and pattern damage for all the emerged adults (Table 2). In general, cauterization in a given segment corresponds in the adult with growth damages in the tergite of the same number. However, there are two important exceptions. Cauterization of the segment VIII in females and segments VII and VIII in males does not produce defective tergites. Segment VIII is known to have no equivalent tergite in the adult. It corresponds to the genital arch in both males and females, which derives from the genital imaginal disc (Ferris, 1950). In fact, cauterization in the VIIIth segment, although not producing growth damage, very frequently alters the rotation of the adult genitalia in both males and females.

The Vllth tergite is absent in the male, and the male Vlth tergite is thought to derive from the fusion of histoblasts of the Vlth and Vllth segments. The following arguments lead to a different conclusion. Cauterization in the Vlth male segment completely suppress the adult tergite in 6 out of 28 cases (21%), as in females (13/49, 27%). Cauterization in the Vllth segment of females leads to complete lack of the Vllth tergite in 64% of the cases (20/31). However, cauterization of the Vllth segment in males did not suppress the Vlth tergite in 18 emerged adults. The slight damage observed in the Vlth tergite possibly corresponds to overlapping damage in neighbour segments (Fig. 5). We conclude that the histoblasts of the male Vllth tergite, if they exist, do not give rise to any visible adult cuticular structures. In order to see the correspondence between the male and female Vlth tergites we studied 184 XC2, wvG/yf36 gynanders with a Vlth and Vllth hemitergite or parts of them of different sex. In all cases the cells differentiating in an entire or reduced Vllth hemitergite were always female. Irrespectively of the presence of a Vllth female hemitergite, the entire Vlth hemitergite was larger when formed by male cells. This was also true when the male tissue of a mosaic Vlth hemitergite occupied the caudal portion of the hemi-tergite. In these cases of mosaic Vlth hemitergites pigmentation was an autonomous feature of the sex of the adult cells (see Dobzhansky, 1931; Stern, 1966). Conversely, female Vlth hemitergites were of normal size and pigmentation even when the Vllth tergite was lacking. Thus, the appearance of a larger Vlth tergite in males and of a Vllth tergite in females depends on the sex of the cells populating the corresponding larval segments. Interestingly, the Vllth segment spiracle appears in the Vlth segment in males but also in females lacking the Vllth tergite. Thus, possibly the appearance of two spiracles in the Vlth male tergite results from the lack of a Vllth tergite rather than from the fusion of tergites VI and VII.

(c) Cauterization in different developmental stages

The data of García-Bellido & Merriam (1971 b) show that the number of histoblasts of each Anlage remains constant throughout the embryonic and larval development. With the onset of metamorphosis, following the formation of the puparium, the presumptive tergite cells start dividing. They spread over the corresponding tergite area until 21–24 h, when divisions cease. Therefore, we studied whether tergite damage varies, following cauterization of the Vth segment, in different developmental stages. The data in Table 3 show that cauterization in regions E and F of the Vth larval segment at the end of the second instar and throughout the third instar leads to similar growth and pattern damages. Thus moulting, apparently, does not correct previous damage in the anlage.

Table 3

Effects of cauterization in regions E or F of the Vth hemisegment of larvae and pupae of different ages

Effects of cauterization in regions E or F of the Vth hemisegment of larvae and pupae of different ages
Effects of cauterization in regions E or F of the Vth hemisegment of larvae and pupae of different ages

Cauterization during the first 12 h after the formation of the puparium causes growth damage in the corresponding adult tergite. Cauterization later leads to lesser damage, and it is ineffective after 20 h. Interestingly, cauterization of pupae older than 12 h in the Vth abdominal segment causes damage in tergites IV, III, and II but no longer in the Vth. This is probably due to the fact that at pupation, which takes place 12 h after pupariation, the abdomen of the pupa shrinks inside of the third instar cuticle, now transformed into the puparium.

III Variations in the number of adult elements

Cauterization of the larval abdominal segments leads to distorted patterns in the corresponding adult tergites. These distortions consist on the lack of cuticular structures such as trichomes, chaetes, lack of pigment, and disarranged distribution of these elements within the adult tergite.

This pattern damage could result from the selective killing of presumptive cells for one or several of the adult cuticular elements. This assumption does apparently not hold, since clones induced during the larval period contain all combinations of different adult elements, which indicates that larval histoblasts have not determined cell lineages (García-Bellido & Merriam, 1971 b). Alternatively, pattern damage might result (1) from damage on the substratum on which the histoblasts proliferate and locate during the population of the presumptive tergite area, (2) from an insufficient number of remnant imaginal cells or an inadequate cell density to create organized and complete patterns. In order to test these possibilities we will analyse the different pattern damages with respect to (1) the cauterized larval region, (2) the location of the remnants on the presumptive tergite area, (3) the size of the remnant tergite, and (4) the relations between the different cuticular elements themselves.

Let us first analyse the relationships between the cauterized larval region and the size and location of the adult remnant tergite. Cauterization in different regions of the larval segment leads to either total lack, to partial development or to complete formation of the adult hemitergite. However, as seen before (Fig. 4), remants, of 1–8/9 the size of the adult hemitergite, appear following cauterization in any larval region. Moreover, there is no clear correlation between the cauterized region and the location of the remnant in the adult tergite area. In fact, although the regions E and F are the most sensitive to histoblast killing, remnants are preferentially located in these regions in the adult. Conversely, cauterization in regions A, D and G, which is poorly effective, inducing growth damage, leads to incomplete tergites in which the corresponding adult regions are frequently unoccupied. Fig. 3 (b) presents the different adult regions classified according to the frequency in which they are occupied by remnants. The shape of somatic recombination clones depends on the location in the tergite area, being elongated in the middle and concentric in the margins (Löbbecke, 1958; García- Bellido & Merriam, 1971 b). This shape suggest a centrifugal proliferation from the original histoblast nest. Similarly, a comparison of Figs. 3(b) and (a) suggests that the histoblasts proliferate and expand from the region where they lie in the larva to the adult hemitergite margins. As seen before, regeneration is low and the killing of some of the histoblasts by cauterization prevents the total population of the presumptive hemitergite. These are precisely the marginal regions which remain unoccupied under these circumstances. The tergite remnants are characterized by having trichomes on a hard and slightly pigmented cuticle, similar to the majority of the adult tergite. Parallel to the margins of the remnants the cuticle has ridges and the extant trichomes are orientated more or less perpendicular to them. The zone surrounding these remnants is always covered by a thin transparent cuticle without trichomes or pigment. This wound cuticle occupies the entire presumptive area in the naked hemitergite. Its origin is unknown (but see Discussion).

Since there is no positive correlation between the region of cauterization in the larva and the region occupied by the tergite remnants, we will try to correlate the presence of the different cuticular elements with both the size and the location of these defective tergites. The different cuticular elements appear more frequently the larger is the remnant (Fig. 6). As mentioned before, remnants always have trichomes and the general yellowish pigmentation. We find micro-chaetes in some 1 /9 remnants, but they appear in all the remnants when their size reaches a surface corresponding to 6/9 of the tergite. Similarly, the mean number of microchaetes per remnant increases, although not proportionally, to the increase in size. Since the 1/9-6/9 remnants occupy preferentially the central regions of the tergite they should have always microchaetes. Thus, the appearance of microchaetes is not simply correlated with either the size or the location of the tergite remnant.

Fig. 6

Relationship between the size of the tergite remnant and the appearance of the different cuticular elements. Each point represents the percentage of tergites with a given type of element. ○ Microchaetes; □ macrochaetes; △, dark pigment; the mean number of such elements per tergite is presented in numerals.

Fig. 6

Relationship between the size of the tergite remnant and the appearance of the different cuticular elements. Each point represents the percentage of tergites with a given type of element. ○ Microchaetes; □ macrochaetes; △, dark pigment; the mean number of such elements per tergite is presented in numerals.

The same situation applies to the appearance of macrochaetes and of the posterior band of pigment in tergite remnants (Fig. 6). Macrochaetes and pigment only appear when the remnant occupies the posterior part of the tergite. But again this apparently is not the casual condition, for remnant tergites covering completely the posterior part of the tergite may not include macrochaetes or pigment. The posterior band of pigment and the macrochaetes appear only in remnants which have also macrochaetes, 9·6% have microchaetes and pigment and 22-9% have microchaetes, pigment and macrochaetes. In some remnants we found pigment without macrochaetes. Thus, the appearance of macrochaetes depends directly on the existence of pigment, or both are causally dependent on the same determinants.

All the previous findings taken together suggest that there is no direct causal relationships between the location of the tergite remnant or its size and the appearance of any given cuticular element. That the size of the remnant does not determine alone the appearance of these elements is further confirmed by the existence of complete tergites with no chaetes or with more chaetes than normal. Entire hemitergites having less than 20 microchaetes or more than 60 on the one hand and with less than 2 macrochaetes and more than 9 on the other, occur in about a 5% of the tergites. The extreme numbers found were 3 and 247 microchaetes and 0 and 14 macrochaetes, figures differing highly from the normal amount of 46 and 8. Interestingly, these abnormalities occur following cauterization in marginal regions. Cauterization in the caudal margin leads to an increase in the number of microchaetes and to a reduction on the number of macrochaetes. However, cauterization in other marginal regions leads to an increase in both microchaetes and macrochaetes. In all the hemitergites with less or more chaetes than normal, the chaetes appear at regular distances from each other, filling the available hemitergite surface.

IV Variations in the organization of the adult elements

Cauterization in certain regions of the larval abdominal segment leads to changes in the organization of the tergite elements. These changes may be restricted to the cauterized hemitergite or involve neighbouring hemitergites. The majority of the obtained abnormalities correspond to these already described by Maas (1948), Sobels (1952), Zimmermann (1954) and Löbbecke (1958).

The most interesting pattern damage restricted to one cauterized hemitergite consists in anterior-posterior duplications. These duplications correspond to the appearance of the typical pattern of heavy pigment and macrochaetes of the caudal margin in the cephalad margin of the hemitergite (Fig. 7 b). Machrochaetes and microchaetes in the anterior part are orientated cephalad. The line of symmetry in these duplications is the medial transverse line of the hemitergite. The anterior pigmented band does not necessarily contact with the posterior band; when it does, it always is along the ventral side of the hemitergite. Duplications are not very frequent: we found 16 cases and only in remnants larger than 6/9. Interestingly, these duplications result from cauterization in region B (4 cases) and C (12 cases) only. The reciprocal duplication, namely the formation of the anterior pattern in the posterior margin, was never detected.

Fig. 7

Schematic representation of the most frequent forms of pattern damage found after cauterization, (a) Contralateral interruption, (b) anterior-posterior duplication, (c, d) contralateral fusions, (e, f) colateral fusions, (g, h) colateral penetrations (description in text).

Fig. 7

Schematic representation of the most frequent forms of pattern damage found after cauterization, (a) Contralateral interruption, (b) anterior-posterior duplication, (c, d) contralateral fusions, (e, f) colateral fusions, (g, h) colateral penetrations (description in text).

Other pattern damages involve the treated hemitergite as well as the neighbouring ones. Following cauterization of certain larval regions, fused hemitergites, invasions of cells of neighbouring colateral and contralateral hemitergites often appear (Table 4).

Table 4

Intertergital pattern damage following cauterization in the different regions of the larval hemisegment

Intertergital pattern damage following cauterization in the different regions of the larval hemisegment
Intertergital pattern damage following cauterization in the different regions of the larval hemisegment

The most frequent pattern damage between contralateral hemitergites consists in the incomplete contact between the left and right hemitergites of the same segment (Fig. 7 a). When following cauterization the left hemitergite is absent, the right hemitergite does not completely cover its presumptive area, but grows short of the medial line. In other cases more or less complete left hemitergites may contact with the right ones along a fraction of the medial line only (Fig. 7 a). These abnormalities were found in 152 otherwise-entire tergites (27·7%). They result from cauterization in any larval region, but the most important fraction results from cauterization in regions A, D, G and J - that is, in regions lying in the medial or dorsal part of the larval segment (Table 4). Since these regions correspond to the adult ones which show the pattern-abnormality possibly these abnormalities result from damage of the dorsal separation line in which proliferating histoblasts of contralateral hemitergites make contact.

Other induced pattern-damages consist in fusion between neighbouring hemitergites. Fig. 7 presents the various types of fusions we found in the adult tergite following cauterization in different regions of the larval Vth segment. We subdivided these intertergital fusions into two main groups.

A first group corresponds to contralateral fusions between hemitergites of different segments - that is, between either IVth and Vth (type I) or Vth and Vlth (type II) segments. It is characteristic of these fusions to maintain all the elements and the general normal pattern; they furthermore appear in almost entire hemitergites. Following fusion of contralateral segments the pigmented band of the anterior tergite may contact with the pigmented band of the posterior . one and then the present macrochaetes follow the posterior margin of both hemitergites (Fig. 7c). In another fusion type (Fig. 7d) we observe the same pattern except that the pigmented band of the anterior hemitergite does not join the pigmented band of the posterior hemitergite. They are characteristic of cauterization in the hemisegment regions lying most dorsal or medial in the larva. Cauterization in the anterior regions A, B, D and G leads to type I fusions and cauterization of posterior regions; G and J leads to type II fusions.

A second group includes colateral fusions between IVth and Vth (type 1) and Vlth and Vth (type II) tergites. We have found many different kinds of colateral fusions. Fusions restricted to the central part of the hemitergite margin have characteristically conflicting orientations of the trichomes and chaetes of both fused segments (14 cases). The trichomes and chaetes of the posterior segment are directed cephalad and the macrochaetes and pigmented band of the anterior segment consistently disappear (Fig. 7e). Such changes in orientation are similar to the disarrangements which appear following experimental removal of the intersegmental regions in other insects (see Lawrence, 1970). When the fusion is continuous along the anterior margin (13 cases) the orientation of trichomes and chaetes is always caudal wards as normal, in both hemitergites. However, the pigmented band and macrochaetes normally disappear (Fig. 7f). Only in three cases have we found, fusion notwithstanding, that the posterior pigmented band and the macrochaetes of the anterior segment remain in place.

The colateral fusion between the Vlth and Vth tergite (type II), following cauterization of the Vth segment, present the same patterns as in the type I. As expected, small fragments of the Vth tergite attached to the Vlth lack both macrochaetes and pigment. These colateral fusions appear following cauterization of the most anterior (A, B and C) and the most posterior (I, J and L) regions of the larval segment (Table 4). Cauterization of the anterior regions leads to type I and of posterior regions to type II pattern alterations. Apparently, these alterations correspond to damage in the larval regions responsible for the separation between neighbouring segments.

Another group of pattern damage corresponds to colateral penetrations of neighbouring hemitergites. In some cases in which the hemitergites are reduced to remnants, or entirely absent, the neighbouring segments may expand and occupy part of the lacking hemitergite. Colateral penetrations can be mistaken for fusions of an entire IVth tergite with remnants of the Vth. However, the 17 cases presented in Table 4 correspond to small shifts of the margin beyond the border line of the normal tergite. This slight shift is enough to produce considerable changes, especially in cases of penetrations of the IVth into the Vth (type 1) (Fig. 7g). The pigment and macrochaetes of the IVth tergite disappear altogether. Penetrations of the Vlth into the Vth segment (type II) leave the pattern of the Vlth tergite unchanged (Fig. 7h). In general, cauterization leading to total killing of the histoblasts is not a sufficient condition for colateral penetrations. Among the 143 cases of total hemitergite loss only 17 cases of penetration have been detected. These correspond to cauterization in anterior, ventral and posterior margin regions, the former leading to type I penetrations and the latter to type II penetrations.

A common feature of the described pattern damages is their correlation with cauterization in the corresponding margin regions, i.e. in the regions delimiting neighbouring larval hemisegments.

The adult hemitergites of Drosophila are formed by the differentiation of imaginal cells deriving from groups of histoblasts, homologous to imaginal discs, situated, segmentally at both sides of the medial dorsal line (Robertson, 1936). Analysis of gynandromorphs and of genetic mosaics has shown that each histoblast group includes about eight cells, set aside from the remaining epidermal cells of the abdomen in the blastoderm stage, which do not proliferate throughout the larval period up to pupariation (García-Bellido & Merriam, 1971 b). Lobbecke (1958) reported a slight proliferation during the larval period. The histoblasts have not yet been detected histologically in the larva. Indirect analysis has suggested that they are scattered among the larval epidermal cells or clustered in two main groups, one anterior and one posterior, within each hemi-segment (Lobbecke, 1958). However, the described cauterization experiment; show that they are clustered possibly in a single group located dorsally in the middle of each larval hemisegment, between two consecutive rows of hooks (Fig. 3). Thus, the adult tergites in Drosophila have a different origin from those in other insects, such as Rhodnius (Wigglesworth, 1940; Locke, 1959, 1960), Oncopeltus (Lawrence, 1966) and Galleria (Piepho, 1955; Marcus, 1962; Stumpf, 1965), in which important morphogenetic analysis of the tergite has been carried out. In these insects the cuticular structures of the adult tergite derive from the same epidermal cells of the larva.

Cauterization experiments in various regions of the larval segment have shown that the proliferative growth of the histoblasts occurs concentrically from the primitive location. When cauterization kills some of the primitive histoblasts the remaining cells only poorly compensate that loss and the adult tergite is reduced in size (Fig. 3 b). It is difficult to explain this lack of regenerative regulation. It could be due to either a fixed determination in the number of presumptive divisions of each histoblast or to a fixed mitotic rate, which would lead to an interruption of growth when the hormonal conditions for cell differentiation begin. Apparently, different mechanisms should account for the formation of the wound cuticle in the naked hemitergite regions (see page 407). As in the case of wound cuticle in Rhodnius (Wigglesworth, 1940), this cuticle could derive from migration from neighbouring epidermal cells. Cells deriving from the ventral region occasionally penetrate into the tergite area, but they differentiate the typical cuticular structures of the ventral side of the abdomen. Cauterization resulting in the lack of the entire hemitergite produces in some cases the lack of the corresponding spiracles and yet the wound cuticle is still formed. From histological data Robertson (1936) has shown that the spiracles derive from an Anlage different than that of the tergites or sternites. Thus, in the present case, the wound cuticle may derive from cells of the intertergital region of the neighbouring tergites, which are unpigmented and do not differentiate trichomes.

The border of each hemitergite apparently does not result from the mutual limitation in the spreading of the proliferating histoblasts of neighbouring hemi-segments. Abnormal connexions and penetrations between neighbouring hemi-tergites occur in otherwise normal hemitergites, which indicates that they do not result from direct damage to the histoblasts. In fact, these pattern abnormalities result from cauterization in marginal regions of the larval hemisegment, far separated from the location of the histoblast nest. Cauterization of the dorsal medial line leads frequently to interruptions in the fusion of contralateral hemi-tergites and also to fusion of contralateral hemitergites of adjacent segments. Cauterization in the anterior margin of the larval segment leads to fusions of adjacent colateral tergites and similarly occurs after cauterization of the posterior margin. Moreover, when cauterization in the anterior or posterior margin leads simultaneously to the killing of the histoblasts, the adjacent histoblasts grow over and partially penetrate the presumptive area of the lacking hemitergite (Fig. 7). We assume that the effect of cauterization consists in the damage or blockage of a margin signal of the substratum upon which the histo-blasts proliferate. This margin signal appears to be qualitatively different in the anterior and dorsal margins.

The nature of these marginal signals is unknown. Margin properties have been found in other insects in transplantation and rotation experiments with pieces of tergite integument (Piepho, 1955; Locke, 1959, 1960; see Lawrence, 1970). In these cases the margin characteristics seem to exist in the cells themselves in the form of a gradient with a maximal low and high level in either the anterior or the posterior margins. Since in Drosophila the tergites originate from histoblasts spreading upon the larval substratum, these margin signals should arise in either the internal surface of the larval cuticle, by a mechanism similar to ‘contact guidance’ (Weiss, 1941), or in the larval epidermal cells - before or during their histolysis - by cell-to-cell induction or as a response to short-range diffusible substances. The role of the larval substratum is very conspicuous in those cases where the abnormal connexion between the adult hemitergites corresponds with the abnormal connexion of the larval hemisegments (Sobels, 1952; Lbbbecke, 1958).

As in other insects, the adult cuticular structures of the Drosophila tergites are orientated caudalwards. Although the population of the tergite occurs radially from the place of the Anlage location, the orientation of the adult cuticular elements do not follow these migration lines. This is not surprising since after a wound the surrounding epidermal cells migrate to form a wound cuticle and lose their previous orientation, although maintaining other determined characters (Wigglesworth, 1940). This caudal orientation is changed under experimental conditions. In tergite remnants which do not reach the presumptive margins, the trichomes and chaetes are directed perpendicularly to the actual margins. This situation is more extreme in anterior-posterior duplications (Fig. 7 b), or in contralateral fusions between adjacent segments, where the chaetes are orientated perpendicularly to the connecting margin - that is, at different angles with the body axis (Fig. 7 c). In colateral fusions conflicting orientations are very conspicuous (Fig. 7 e) and correspond to those found in other insects following interruption of the intersegmental membrane (Locke, 1960; Lawrence, 1966). These conflicting orientations could result from the tendencies of the cuticular elements to orientate perpendicularly to the new margins, perhaps maintaining their migrational orientation, or alternatively reflect the reaction of the cells to a gradient, which originates in the margin, as suggested by rotation experiments in the tergite epidemis of other insects (see Lawrence, 1970).

The number of elements of the adult pattern seems to be controlled. Chaetes appear in tergite remnants only when they reach a minimal size which is larger than that occupied by a chaete in the normal hemitergite. However, a minimal tergite size is not the limiting factor, since cauterized entire tergites may have several times less or more chaetes than normal. We do not know which are the conditions determining the number of chaetes. Perhaps it is interesting to note that these abnormalities appear after cauterization in certain marginal regions of the larval segment.

Signals of the substratum seem to be required for the appearance of certain cuticular structures. Clonal analysis has shown that the adult cuticular elements of the pattern, microchaetes, macrochaetes and pigment are not clonally determined in the histoblasts of the larval Anlage (García-Bellido & Merriam, 1971). The determination to become one of these elements must occur during proliferation. The different elements of the adult cuticle appear to be, in part, independently determined (p. 408). Microchaetes appear when the remnants occupy the central portion of the tergite area; macrochaetes and the band of heavy pigment only appear when the remnant includes the majority of the posterior margin. In hemitergites which do not cover this margin or, interestingly, over-pass it by fusion (Fig. 7e,f’) or by penetration (Fig. 7g, h), macrochaetes and pigment are lacking altogether. Thus two conditions seem to be required for a hemitergite differentiating these two elements; the histoblasts have to occupy the correct region and this region has to be tergital margin. Occasionally the posterior band of pigment and macrochaetes appear also in the anterior margin of the hemitergite. Such duplications (Fig. 1b) were also found in Drosophila (Zimmermann, 1954; Löbbecke, 1958) and in Oncopeltus (Löbbecke in Fig. 9, 1970). The interpretation of this abnormality put forward by Löbbecke (1958), that it could derive from regeneration of already determined histoblasts of the posterior part moving into the anterior, apparently does not hold, in view of the previous discussion. The appearance of a new posterior margin in the anterior region could again be due to the damage of the substratum. In fact, these duplications only appear following cauterization in B and C; that is, in anterior regions.

The final arrangements of chaetes into a pattern may result from more complicated mechanisms. The distances between the chaetes are more or less regular, independently of the number of chaetes per hemitergite. Wigglesworth (1940) explains bristle location in Rhodnius tergites as resulting from the competition of the epidermal cells for diffusible substances. This interpretation does not simply explain other experimental findings in Drosophila. Clonal analysis has shown that presumptive chaete-forming cells, but not trichome cells, are singled out from their clones, the marked chaetes appearing located at regular distances with respect to the other chaetes (García-Bellido & Merriam, 1971 b). Thus, previous determination and subsequent migration could play a rule in the mechanism determining the final pattern of the tergites.

La cautérisation dans différentes régions d’un segment larvaire cause divers effets sur la croissance et l’organisation de l’hémitergite adulte. On a vérifié une correspondance entre tous les segments de la larve et de l’adulte, exception faite, chez le mâle, du septième segment abdominal, lequel n’a pas, apparemment, de représentation cuticulaire adulte. Les histoblastes qui forment l’ébauche imagínale sont graupés dans la région centrale de l’hémisegment larvaire. Après cautérisation, les histoblastes viables ne peuvent remplacer complètement ceux qui ont été tués, donnant lieu à des tergites incomplets.

Des altérations caractéristiques du ‘pattern’ cuticulaire suivent la cautérisation de certaines régions marginales du segment larvaire. Ainsi, on provoque, même en ne tuant pas les histo-blastes, des fusions et pénétrations spécifiques entre les hémiterguites collatéraux et contralatéraux voisins. On discute les effets de la cautérisation sur la détermination du nombre des éléments cuticulaires et sur leur disposition dans le tergite adulte.

We thank Dr P. Lawrence for his critical reading of the manuscript and G. Morata and P. Ripoll for their constructive discussion.

Becker
,
H. J.
(
1957
).
Über Röntgenmosaikflecken und Defektmutationen am Auge von Drosophila und die Entwicklungsphysiologie des Auges
.
Z. Vererb Lehere
88
,
333
373
.
Bryant
,
P. J.
(
1970
).
Ceil lineage relationships in the imaginai wing disc of Drosophila melano-gaster
.
Devi Biol
.
22
,
389
411
.
Bryant
,
P. J.
&
Schneiderman
,
H. A.
(
1969
).
Cell lineage, growth and determination in the imaginai leg disc of Drosophila melanogaster
.
Devi Biol
.
20
,
263
290
.
Dobzhansky
,
T.
(
1931
).
Interaction between female and male parts in gynandromorphs of Drosophila simulaes
.
Wilhelm Roux Arch. EntwMech. Org
.
123
,
719
746
.
Ferris
,
G. F.
(
1950
).
External morphology of the adult
.
In Biology of Drosophila
(ed.
M.
Demerec
), pp.
268
418
.
New York
:
Wiley
.
García-Bellido
,
A.
(
1968
).
Cell lineage in the wing disc of Drosophila melanogaster
.
Genetics
60
,
181
.
García-Bellido
,
A.
(
1972
).
Some parameters of mitotic recombination in Drosophila melanogaster
.
Mol. Gen. Genetics
115
,
54
72
.
García-Bellido
,
A.
&
Merriam
,
J. R.
(
1971a
).
Parameters of the wing imaginai disc development of Drosophila melanogaster
.
Devi Biol
.
24
,
61
67
.
García-Bellido
,
A.
&
Merriam
,
J. R.
(
1971b
).
Clonal parameters of tergite development in Drosophila melanogaster
.
Devi Biol
.
26
,
264
276
.
Lawrence
,
P. A.
(
1966
).
Gradients in the insect segment: the orientation of hairs in the milkweed bug, Oncopeltus fasciatus
.
J. exp. Biol
.
44
,
607
620
.
Lawrence
,
P. A.
(
1970
).
Polarity and patterns in the postembryonic development of insects
.
Adv. Insect. Physiol
.
7
,
197
266
.
Löbbecke
,
E. A.
(
1958
).
Über die Entwicklung der imaginalen Epidermis des Abdomens von Drosophila, ihre Segmentierung und die Determination der Tergite
.
Biol. Zbl
.
77
,
209
237
.
Locke
,
M.
(
1959
).
The cuticular pattern in an insect, Rhodniusprolixus Stâl
.
J. exp. Biol
.
36
,
459
477
.
Locke
,
M.
(
1960
).
The cuticular patterns in an insect-the intersegmental membranes
.
J exp. Biol
.
37
,
398
406
.
Maas
,
A. H.
(
1948
).
Über die Auslösbarkeit von Temperaturmodifikationen wâhrend der Embryonalentwicklung von Drosophila melanogaster, Meigen
.
Wilhelm Roux Arch. Entw-Mech. Org
.
143
,
515
572
.
Marcus
,
W.
(
1962
).
Untersuchungen über die Polaritât der Rumpfhaut von Schmetterlingen
.
Wilhelm Roux Arch. EntwMech. Org
.
154
,
56
102
.
Piepho
,
H.
(
1955
).
Über die Ausrichtung der Schuppenbalge und Schuppen am Schmetter-lingsrumpf
.
Naturwissenschaften
42
,
22
.
Postlethwait
,
H. H.
&
Schneiderman
,
H. A.
(
1971
).
A clonal analysis of development in Drosophila melanogaster’. Morphogenesis determination and growth in the wild type antenna
.
Devl Biol
.
24
,
477
519
.
Robertson
,
CH. W.
(
1936
).
The metamorphosis of Drosophila melanogaster, including an accurately timed account of the principal morphological changes
.
J. Morph
.
59
,
351
399
.
Sobels
,
F. H.
(
1952
).
Genetics and morphology of the genotype ‘asymmetric’ with special reference to its ‘abnormal abdomen’ character (Drosophila melanogaster)
.
Genética
26
,
117
279
.
Stern
,
C.
(
1966
).
Pigmentation mosaicism in intersexes of Drosophila
.
Rev. suisse Zool
.
73
,
339
355
.
Stumpf
,
H.
(
1965
).
Deutung der Richtungsmuster der Schuppen von Galleria mallonella auf Grund eines Konzentrationsgefâlles
.
Naturwissenschaften
52
,
522
.
Wigglesworth
,
V. B.
(
1940
).
Local and general factors in the development of ‘pattern’ in Rhodnius prolixus (Hemiptera)
.
J. exp. Biol
.
17
,
180
200
.
Weiss
,
P.
(
1941
).
Nerve patterns: The mechanics of nerve growth
.
Growth Suppl
.
5
,
163
203
.
Zimmermann
,
W.
(
1954
).
Über genetisch und modifikatorisch bedingte Storungen der Segmentierung bei Drosophila melanogaster
.
Z. VererbLehre
86
,
327
372
.