The distribution of regulative potential was investigated in the wing disc of Drosophila. Ten complementary pairs of fragments were tested for their capacity to regenerate or duplicate. The distribution of positional values resulting from this data was found to be very unequal; six of the twelve clockface values were tightly clustered round the anterior-posterior compartment boundary. Despite this, complementarity between regeneration and duplication was generally maintained.

One of the most striking results to come out of Drosophila disc regeneration work is that of two complementary fragments, one duplicates existing structures while the other regenerates (for review see Bryant, 1978). Usually the smaller fragment duplicates, but this is not always the case; for example the upper medial quarter of the leg disc regenerates while the remaining three-quarters duplicates (Schubiger, 1971). In the terms of positional information (Wolpert, 1971) and the polar coordinate model (French, Bryant & Bryant, 1976), this quarter has over half the positional values. This uneven spacing of values has been found in the haltere disc (van der Meer & Ouweneel, 1974) as well as the leg disc, and inferred from indirect evidence in axolotl and newt limbs (French et al., 1976; Stocum, 1978; Holder, Tank & Bryant, 1980). However in the wing disc, the evidence has suggested an even spacing (Bryant, 1975). The present experiments were undertaken to find whether this was in fact the case.

Host flies were of ebony11 genotype; donors of ebony11 or occasionally yellow; multiple wing hairs genotype. For descriptions of these mutants see Lindsley and Grell (1968). Flies were raised at 25 °C on standard cornmeal/ syrup/agar medium seeded with live yeast. Wing discs were removed from late-third-instar larvae in insect Ringer. Fragments were cut with tungsten needles and implanted into the body cavities of well-fed 1-to 3-day-old fertilized adult females where they remained for 5–7 days. They were then removed and reimplanted into the body cavities of late-third-instar larvae using the method of Ephrussi and Beadle (Ursprung, 1967). When the hosts emerged as adults, the metamorphosed implants were removed, mounted in Hydramount, and scored for the cuticular markers shown in Fig. 1. These were identified with the aid of the descriptions of Bryant (1975).

Fig. 1.

Fate map of the wing disc (after Bryant, 1975). Presumptive wing blade is stippled. Compartment boundaries are shown in broken lines. The position of the dorsal end of the anterior-posterior boundary was taken from data of Garcia-Bellido, Ripoll & Morata (1976) and of the ventral end from that of Adler (1978).

Fig. 1.

Fate map of the wing disc (after Bryant, 1975). Presumptive wing blade is stippled. Compartment boundaries are shown in broken lines. The position of the dorsal end of the anterior-posterior boundary was taken from data of Garcia-Bellido, Ripoll & Morata (1976) and of the ventral end from that of Adler (1978).

In all the experiments reported here, the wing disc was cut into two fragments, both of which were tested for their ability to regenerate. The cuts used are shown in Figs 2 and 3. Testing of both fragments was done for two reasons. Firstly, to make sure that regeneration and duplication were indeed complementary ; the frequency of regeneration in one fragment of a pair should equal the frequency of duplication in the other. This was generally found to be true (Table 1). Secondly, to check the accuracy of cutting; Bryant (1975) has shown that duplicating fragments have all the structures, and only those structures, expected from the fate map. If both fragments of a pair are tested, and one regenerates while the other duplicates, the position of the cut can be checked by consideration of the structures present in implants of the duplicating fragment. If both of a pair regenerate sometimes and duplicate sometimes, the duplicating implants of both can be used for this purpose. Duplicating implants were found to have those structures expected from the fate map of Bryant (1975), showing that the position of the cuts coincided with their intended location.

Table 1.

Regeneration and duplication frequencies for the fragments shown in Figs. 2 and 3 

Regeneration and duplication frequencies for the fragments shown in Figs. 2 and 3
Regeneration and duplication frequencies for the fragments shown in Figs. 2 and 3
Fig. 2.

(a) The pairs of fragments used in the first set of experiments. All fragments regenerated at a frequency of 19% or over; lines I–V are therefore lines of transition between regeneration and duplication (see Results) and must have 6 values on either side, (b) The distribution of values resulting from these data. Values 2, 10 and 11 were placed equidistant from their neighbours.

Fig. 2.

(a) The pairs of fragments used in the first set of experiments. All fragments regenerated at a frequency of 19% or over; lines I–V are therefore lines of transition between regeneration and duplication (see Results) and must have 6 values on either side, (b) The distribution of values resulting from these data. Values 2, 10 and 11 were placed equidistant from their neighbours.

Fig. 3.

(a) The pairs of fragments used in the second set of experiments, (b) The final allocation of values.

Fig. 3.

(a) The pairs of fragments used in the second set of experiments, (b) The final allocation of values.

Criterion for regeneration

An implant was considered to have regenerated if it contained at least one structure (other than notum and scutellum, see below) which is not expected from the fate map and lies at some distance from the cut edge. The structures used as criteria for regeneration are in bold in Tables 2 and 3.

Table 2.

Structures differentiated by implants of the fragments shown in Fig. 2, cultured in adult females for 5–7 days before transfer to larval hosts for metamorphosis

Structures differentiated by implants of the fragments shown in Fig. 2, cultured in adult females for 5–7 days before transfer to larval hosts for metamorphosis
Structures differentiated by implants of the fragments shown in Fig. 2, cultured in adult females for 5–7 days before transfer to larval hosts for metamorphosis
Table 3.

Structures differentiated by implants of the fragments shown in Fig. 3, cultured in adult females for 5–7 days before transfer to larval hosts for metamorphosis

Structures differentiated by implants of the fragments shown in Fig. 3, cultured in adult females for 5–7 days before transfer to larval hosts for metamorphosis
Structures differentiated by implants of the fragments shown in Fig. 3, cultured in adult females for 5–7 days before transfer to larval hosts for metamorphosis

Complementarity between regeneration and duplication

In general it was easy to decide whether a particular implant had regenerated or duplicated; implants considered to have regenerated by the above criterion rarely had duplicated structures. If they did, in the majority of implants there was a single duplicated structure close to the line of cutting, a phenomenon which can be explained as due to anomalous healing (Reinhardt, Hodgkin & Bryant, 1977). There were however three cases in which regeneration of one structure was accompanied by duplication of two or more. The most important of these was the tendency of presumptive ventral tissue to produce ‘adventitious bristles’ (Bryant, 1975) while duplicating; while in most implants these bristles were few in number (one to ten) and unidentifiable, occasionally there were many and it was quite clear that they were from the notum. In other words, duplicating ventral fragments which lack notum have the ability to regenerate it; this phenomenon is at present under investigation (Karlsson & Smith, in preparation). Structures other than notum and scutellum appear rarely and only after much longer culture periods than used in the present experiments. Notum and scutellum however appeared at high frequency; this meant that neither of these structures could be used as criteria for regeneration. They were scored in the following way. If an implant had notum or scutellum in its fate map, or had regenerated according to other criteria, then these structures were scored as notum and scutellum. If the implant did not include notum or scutellum in its fate map and had formed no other structures well outside the fate map, notum-like bristles were scored as ‘adventitious bristles’.

The other two cases of simultaneous regeneration and duplication occurred in fragments la and IVa. 4/18, and 4/30 implants respectively, duplicated several existing structures and regenerated another; this regenerated structure was yellow club in fragment la and alar lobe in fragment IVa.

In these three cases, a decision had to be made about whether to consider the anomalous implants as having regenerated. It was decided not to do so if the regenerated structure was not always clearly identifiable. Thus duplicating ventral fragments producing ‘adventitious bristles’ were considered to have duplicated, even in those cases in which the bristles were clearly from the notum. The eight anomalous implants of fragments la and IVa, on the other hand, were included in the regenerating category.

A second type of deviation from the rule of complementary between regeneration and duplication can be seen in Table 1. Here the regeneration frequency of one fragment of a pair has been subtracted from the duplication frequency of the other. Since each fragment can regenerate or duplicate (or both), the 10 pairs produced 20 such calculations. For 15 of these this figure is less than 10%, showing that complementarity is in general very well maintained. Of the remaining five cases, there are two (pairs II and VII) in which one of the pair is predominantly posterior and shows a high frequency of duplicated structures. This high frequency seems to be characteristic of posterior fragments (Karlsson, unpublished results) and is conceivably due to their having a high concentration of structures in which duplication is easy to detect.

Two of the remaining cases concern pair III, both fragments of which appear to have low regeneration frequencies. That this is not an artifact due to poor growth is shown by the quite high frequency of duplication in both fragments. Nor is it due to a lack of adequate markers for regeneration as both fragments have several excellent ones.

Fragment VIII b, the fifth case, has a rather higher level of duplication than expected from the complete failure of regeneration in its complementary fragment. That this latter fragment is the only one wholly confined to the anterior compartment could conceivably have a bearing on this result.

Allocation of positional values

The paradigm of the polar coordinate model (French et al., 1976) was used, and the twelve circumferential values were allocated on the following basis :

  1. If both fragments of a pair regenerated at a frequency of 20% or over (19 % in the case of fragment lb) then the line separating them was considered to be a line of transition between regeneration and duplication. Both such fragments were allocated 6 of the 12 values, so that intercalary regeneration by the ‘shortest route’ (French et al., 1916) around the disc circumference could either repeat the existing values (duplicate) or replace missing ones (regenerate).

  2. A fragment which never regenerated could not have more than five values.

    The method employed for finding the distribution of values was to find lines of transition between regeneration and duplication at several different angles. This was done for the following reason: any two lines each having an equal number of values on either side make four quadrants of which both opposite pairs must have the same number of values. Thus if several such lines are found, a distribution of values which obeys this rule for all pairs must be the only possible one.

Two transition lines were known from the work of Bryant (1975); these were confirmed (lines IV and V, Fig. 2) and three more were found. Values were allocated by trial and error until a distribution was found in which all fragments had six values, and all pairs of opposite quadrants had the same number of values. Figure 2 shows the cutting lines and the resulting value distribution. Values 2,10 and 11 were placed equidistant from their neighbours; they are so closely spaced that more precise localisation would not be possible within the limits of cutting error. The positions of values 4, 5 and 8 remained to be determined.

The rest of the experiments were done to localise the remaining values and confirm the positions of those already found. Fragments Villa, VIb and Xb were used to allocate values 4, 5 and 8 respectively; none of these three fragments ever regenerated and the three values were therefore placed as shown in Fig. 3b.

Fragment IX b was used to confirm the very close spacing of values 9 to 11. This fragment was cut very close ventrally to fragment Villa; since neither of these two fragments ever regenerated, at least two values must be crammed into this very small space.

The results show that the positional values in the wing disc are very unevenly spaced, half of them being tightly clustered round the anterior-posterior compartment boundary.

The simplest way in which unequal spacing could arise is through differences in cell size or density, closer spacing corresponding to smaller or denser cells, and all cells having the same proportion of the values. However imaginai discs have no such differences which could correspond to differences in value spacing (Ursprung, 1972). Another way is through unequal growth, which could cause an initially equal spacing to become unequal during development. This would provide an explanation for the sparseness of values in most of the posterior compartment; clonal analysis shows that this compartment grows faster than the anterior compartment during at least part of development (Lawrence & Morata, 1976). However if this were the whole story, one would expect clones lying along the anterior-posterior boundary, where values are closest, to be smaller than average, and this is not the case (Garcia-Bellido & Merriam, 1971).

It is possible of course that differences in value spacing have no functional significance whatever, but the very extreme differences found in the present work seem to require an explanation. The spacing of positional values is generally considered to represent the slope of a gradient of some cellular parameter responsible for giving cells information about their position and hence what to do (Wolpert, 1971). If each structure differentiates at a particular threshold value of this gradient, close spacing would mean that thresholds could be closer together; this might be required where particularly complex structures were to differentiate. Indeed, the most complex structures of the wing disc, those of the dorsal and ventral hinges, lie along the anterior-posterior compartment boundary where the spacing of values is closest. Again, though, this cannot be the whole story, as there is no such correlation between value spacing and structure complexity in the leg disc (Schubiger, 1968).

The close spacing at the anterior–posterior compartment boundary, if it is not coincidence, could mean that cells at the boundary have some special property. This could be for example the high point of a gradient; it has been suggested that this boundary represents the common high point of two identical gradients of positional information (Crick & Lawrence, 1975). Supposing that regeneration can only occur down the gradient, fragments lacking boundary would be unable to regenerate, but could only duplicate, as observed. Such a theory would have to explain why many fragments having boundary do not regenerate. The greatest difficulty though would be in explaining why the anterior-posterior boundary is not required for regeneration in the leg disc (Schubiger, 1971; Steiner, 1976).

The presence of the anterior–posterior compartment boundary has been found necessary for distal regeneration to occur in both the wing disc (Wilcox & Smith, 1980; Karlsson, 1980) and the leg disc (Schubiger & Schubiger, 1978), and several observations suggest that something special is happening at compartment boundaries in normal development. Simpson (1976) has found that the phenomenon of cell competition, by which slow-growing cells are eliminated, operates to a much lesser extent at compartment boundaries than elsewhere, suggesting that the rules governing growth are different at the boundary and in the middle. Lawrence & Morata (1976) found that small engrailed clones could cause a shape distortion of the wing blade if they touched the dorso-ventral compartment boundary but not otherwise, and suggested that the boundary was somehow instrumental in controlling growth. It may also be significant that all well-documented compartment boundaries, in all appendages, are aligned along the major growth axes. These observations, together with the finding that the anterior–posterior compartment boundary is important in regeneration, suggest that the significance of compartments may reside in their boundaries, and that it is here that the overall shape of appendages is determined.

It is noteworthy that in spite of wide differences in the spacing of positional values, the complementarity between regeneration and duplication is in general very well maintained. It clearly never happens that both fragments of a pair unequivocally regenerate, or unequivocally duplicate. Whether this applies at the level of individual discs is not known as this would entail fragments from single discs being kept separate, an experiment which has not so far been done.

Complementarity between regeneration and duplication has been elegantly explained by the polar coordinate model (French et al., 1976) in terms of both fragments of a pair having the same positional values at their cut edges and therefore intercalating the same structures. That this is indeed a useful way to view regeneration is confirmed by the internal consistency of the present results, without which it would not have been possible to allocate values.

This work was supported by an MRC Studentship and a Beit Memorial Fellowship.

Adler
,
P. N.
(
1978
).
Positional information in imaginai discs transformed by homeotic mutations
.
Wilhelm Roux’ Arch, devl Biol
.
185
,
271
292
.
Bryant
,
P. J.
(
1975
).
Pattern formation in the imaginai wing disc of Drosophila melanogaster’. fate map, regeneration and duplication
.
J. exp. Zool
.
193
,
49
78
.
Bryant
,
P. J.
(
1978
).
Pattern formation in imaginai discs
. In.
The Genetics and Biology of Drosophila
, vol.
2c
(eds.
M.
Ashburner
&
T. R. F.
Wright
).
London
:
Academic Press
.
Crick
,
F. H. C.
&
Lawrence
,
P. A.
(
1975
).
Compartments and polyclones in insect development
.
Science
189
,
340
347
.
French
,
V.
,
Bryant
,
P. J.
&
Bryant
,
S. V.
(
1976
).
Pattern regulation in epimorphic fields
.
Science
193
,
969
981
.
Garcia-Bellido
,
A.
&
Merriam
,
J. R.
(
1971
).
Parameters of wing imaginai disc development of Drosophila melanogaster
.
Devl Biol
.
24
,
61
87
.
Garcia-Bellido
,
A.
,
Ripoll
,
P.
&
Morata
,
G.
(
1976
).
Developmental compartmentalisation in the dorsal mesothoracic disc of Drosophila
.
Devl Biol
.
48
.,
132
147
.
Holder
,
N.
,
Tank
,
P. W.
and
Bryant
,
S. V.
(
1980
).
Regeneration of symmetrical forelimbs in the axolotl Ambystoma mexicanum
.
Devl Biol
.
74
,
302
314
.
Karlsson
,
J.
(
1980
).
Distal regeneration in proximal fragments of the wing disc of Drosophila
.
J. Embryol. exp. Morph
.
59
,
315
323
.
Lawrence
,
P. A.
&
Morata
,
G.
(
1976
).
The compartment hypothesis
.
In: Insect Development
(ed.
P. A.
Larence
).
Oxford
:
Blackwell Scientific Publications
.
Lindsley
,
D. A.
&
Grell
,
E. H.
(
1968
).
Genetic Variation in Drosophila melanogaster
.
Carnegie Inst. Washington, Publ
. no.
627
.
Reinhardt
,
C. A.
,
Hodgkin
,
N. M.
&
Bryant
,
P. J.
(
1977
).
Wound healing in the imaginai discs of Drosophila. 1. Scanning electron microscopy of normal and healing wing discs
.
Devl Biol
.
60
,
238
257
.
Schubiger
,
G.
(
1968
).
Anlageplan, Determinationszustand und Transdeterminationsleistungen der mannlichen Vorderbeinscheibe von Drosophila melanogaster
.
Wilhelm Roux Arch. EntwMech. Org
.
160
,
9
40
.
Schubiger
,
G.
(
1971
).
Regeneration, duplication and transdetermination in fragments of the leg disc of Drosophila melanogaster
.
Devl Biol
.
26
,
277
295
.
Schubiger
,
G.
&
Schubiger
,
M.
(
1978
).
Distal transformation in Drosophila leg imaginai disc fragments
.
Devl Biol
.
67
,
286
295
.
Simpson
,
P.
(
1976
).
Analysis of the compartments of the wing disc of Drosophila melanogaster mosaic for a temperature sensitive mutant that reduces mitotic rate
.
Devl Biol
.
54
,
100
115
.
Steiner
,
E.
(
1976
).
Establishment of compartments in the developing leg imaginai discs of Drosophila melanogaster
.
Wilhelm Roux’ Arch, devl Biol
.
180
,
9
30
.
Stocum
,
D. L.
(
1978
).
Regeneration of symmetrical hindlimbs in larval salamanders
.
Science
200
,
790
793
.
Ursprung
, &
. (
1967
).
In vivo culture of Drosophila imaginai discs
.
In Methods in Developmental Biology
(eds.
F. H.
Wilt
and
N. K.
Wessels
).
New York
:
Crowell
.
Ursprung
,
H.
(
1972
).
The fine structure of imaginai discs
.
In The Biology of Imaginai Discs
(eds.
H.
Ursprung
and
R.
Nöthiger
).
Berlin and Heidelberg
:
Springer-Verlag
.
van der Meer
,
J. M.
&
Ouweneel
,
W. J.
(
1974
).
Differentiation capacities of the dorsal metathoracic (haltere) disc of Drosophila melanogaster: II. Regeneration and duplication
.
Wilhelm Roux Arch. EntwMech. Org
.
174
,
361
373
.
Wilcox
,
M.
&
Smith
,
R. J.
(
1980
).
Compartments and distal regeneration in the Drosophila imaginai wing disc
.
Wilhelm Roux’ Arch, devl Biol
.
188
,
157
161
.
Wolpert
,
L.
(
1971
).
Positional information and pattern formation
.
In Current Topics in Developmental Biology
6
,
183
224
.
     
  • A, P

    presumptive anterior, posterior

  •  
  • Scu

    Scutellum

  •  
  • HP

    Humeral Plate

  •  
  • UP

    Unnamed Plate

  •  
  • AS 1–3

    Axillary Sclerites 1–3

  •  
  • T Row

    Triple Row of Chaetes

  •  
  • D Row

    Double Row of Chaetes

  •  
  • P Row

    Posterior Row of Hairs

  •  
  • PDR

    Proximal Dorsal Radius

  •  
  • A Lobe

    Alar Lobe

  •  
  • A Cord

    Axillary Cord

  •  
  • YC

    Yellow Club

  •  
  • PVR

    Proximal Ventral Radius

  •  
  • PS

    Pleural Sclerite

  •  
  • PWP

    Pleural Wing Process