Cautery-induced colour patterns in Precis coenia (Lepidoptera: Nymphalidae)

The colour patterns on the wings of Lepidoptera are determined during the prepupal and early pupal stages (Nijhout, 1978, 1984b). Microcautery of the wing epidermis during the period of pattern determination can induce aberrations in the colour pattern, and from such aberrations it has occasionally been possible to deduce the characteristics of the processes underlying pattern determination. Cautery experiments have shown, for instance, that colour pattern determination occurs independently on each wing surface (Nijhout, 1978, 1980a). In the moth, Ephestia kühniella, development of a banding pattern that runs across the forewing has been shown to depend on the activity of three inducing ‘sources’, located on the wing midline (Kühn & Von Engelhardt, 1933; Nijhout, 1978, 1984b). These sources produce a signal that appears to propagate by diffusion. Cauteries of the wing epidermis, placed between a source and the presumptive location of a cross band, cause dramatic alterations in the shape and position of that crossband. Kühn & Von Engelhardt (1933) have proposed that this is due to the fact that dead, cauterized, cells pose an obstacle to the propagation of the inducing signal. In the butterfly, Precis coenia, similar sources, dubbed foci (Nijhout, 1978, 1984b), are responsible for inducing circular eyespot-like patterns on the forewing. When the cells of a focus are killed by cautery before pattern determination begins, the eyespot fails to develop (Nijhout, 1980a). In both these cases the effect of cautery on the morphology of the colour pattern can be attributed straightforwardly to an interference with the normal process of pattern determination: in Ephestia, a local interference with the propagation of an inductive signal, and in the Precis forewing, an actual elimination of the source of the inductive signal.

In contrast to the Precis forewing, where cautery has no effect on the colour pattern unless it is placed on or very near a focus, I have found that a cautery placed almost anywhere on the dorsal hindwing causes the development of a novel colour pattern, identical in its morphology to the eyespots that normally occur on that wing surface. This response to cautery appears to be restricted to the dorsal hindwing of this species. The present paper describes this apparent inductive effect of cautery. It also proposes a physiological mechanism for this form of pattern induction that is not only compatible with the other known effects of cautery, but also helps to explain a peculiar set of pattern aberrations that develop in response to coldshock on the dorsal hindwing of this butterfly.

Larvae of Precis coenia were reared on an artificial diet as described by Nijhout (1980a). Except as noted, cauteries were done on young pupae, 4 to 6h after pupation. Cauteries were made with a sharpened tungsten needle, wrapped to within 2mm from its tip with three loops of nichrome wire and insulated with epoxy glue. Current (DC) was passed through the nichrome wire to heat the tip of the needle to approximately 100 °C. Since in lepidopteran pupae the forewing is firmly glued over most of the hindwing, cauteries of any but the dorsal surface of the forewing must be done blindly by inserting the cautery needle sufficiently deeply to penetrate the desired epidermal layer. As a rule the needle was inserted so as to cauterize both surfaces of both fore- and hindwing. Since the forewing is always folded over the hindwing in the same way, it is possible, after a series of exploratory cauteries, to use landmarks on the forewing cuticle to locate specific areas of the hindwing with reasonable accuracy. For light cauteries the needle was held in place for about 1/2 second. More severe cauteries were accomplished by holding the needle in place for up to 10 seconds. Coldshocks were effected by placing pupae, 4 to 6h after pupation at a temperature of −2 °C for a period of 72 h, as described by Nijhout (1984a).

Computer simulations of diffusion were done by a finite difference method on a 2-dimensional rectangular grid in which any node(s) could be designated as source, sink, or obstacle for the diffusing substance. The diffusing substance propagated with a diffusion coefficient of 8 × 10⁻⁹cm²/sec and decayed with a half life of 12 h. These values were derived as best fits to experimental data on the dynamics of eyespot determination in Precis (Nijhout, 1980a) and have proven to be useful in simulating the development of a variety of patterns (Nijhout, 1984b and in preparation). Diffusion was allowed to continue for 48h, which is the normal period for pattern determination in Precis (Nijhout, 1980a). At that time the gradients produced were nearly at steady state. Sources were modelled as constant-level sources. Concentration of the diffusing substances were in arbitrary units, as described in the legend to the appropriate figures.

Cautery of the dorsal epithelium of the hindwing in pupae of Precis coenia induces the formation of a concentric colour pattern provided the cautery is placed in the presumptive brown ‘background’ area of the wing and provided the age of the pupa is less than 48 h. Figure 1 illustrates the characteristics of the induced pattern. Cauteries of short duration which kill few cells induce a circular patch of black pigment at the site of cautery (Fig. 1A). More severe cauteries induce large patterns that resemble the smaller of the two eyespots on the dorsal hindwing. Induced patterns have an outer black ring and a yellow central field, identical in colour to the black and yellow rings of the normal eyespots. Occasionally the centre of the induced eyespot also bears a patch of orange and black scales, as does the central field of a normal eyespot.

When a cautery is placed close to one of the natural eyespots, the black and yellow rings of the induced patterns fuse with their counterparts in the normal eyespot (Fig. 1C–E). The confluence of the induced pattern with the rings of the natural eyespot is always perfectly smooth, which can be taken as an indication that cautery is somehow either mimicking or evoking the physiological conditions that are normally associated with determination of the outer (peripheral) portions of the eyespot pattern.

Cautery can induce eyespots anywhere in the brown background area of the wing and all points within this area appear to be equally responsive (Fig. 1E). Areas of the hindwing that are overlaid by forewing are no more responsive to cautery than those that are not, so an interaction of fore- and hindwing in the development of supernumerary eyespots is unlikely.

Cauteries placed within the black central disk of one of the natural eyespots on the hindwing either have no obvious effect or cause a diminution in the size of the eyespot. Likewise, cautery on the margin of the wing, within the presumptive area of the submarginal bands and parafocal element (see Nijhout, 1984a,b for nomen-clature and homologies of these pattern elements), causes a local distortion of those patterns but does not induce a novel pattern. When a cautery is placed very near a parafocal element it induces the formation of an incomplete eyespot that is truncated where it meets the parafocal element (Fig. 1F). The simplest explanation for this finding is that the physiological events that determine the parafocal elements and submarginal bands can override pattern induction by cautery. It is unlikely that determination of the parafocal element occurred before cautery and therefore precluded redetermination by cautery, because cautery can induce considerable distortions in the shape of the parafocal elements, and evidently interferes with their proper determination.

I have demonstrated elsewhere that cautery of the foci on the dorsal forewing can abolish the development of the large eyespot on that wing surface (Nijhout, 1980a). Cauteries elsewhere on the dorsal forewing as a rule have no effect on the pattern. Among the nearly 1000 specimens whose forewings have been cauterized during the past 5 years I have encountered only four cases in which cautery induced an eyespot-like pattern but these were not nearly as well differentiated as those illustrated above for the hindwing, or as those that form around transplanted foci (see Nijhout, 1980a). When cauteries are placed on or near the outer dark ring of the large eyespot on the forewing, however, the shape of the outer band often becomes distorted so as to enclose the site of cautery, and the central black disk of the eyespot occasionally develops a bulge in the direction of the site of cautery. Both these distortions of the eyespot are more evident and more readily induced on the ventral surface of the forewing and are illustrated in Fig. 2. Cautery elsewhere on the dorsal or ventral surface of the forewing has no effect on pattern development.

Cauteries of the ventral surface of the hindwing induce substantial distortions of the colour pattern but do not induce eyespot-like patterns. Cautery-induced pattern modification on the ventral hindwing can be quite complex and will be described at a future occasion.

Development of an eyespot-like pattern around the site of cautery thus appears to be an invariable characteristic of the dorsal hindwing. The cautery-induced distortions of the eyespots on the forewing may be low-level manifestations of the same phenomenon, somewhat more readily expressed on the ventral than on the dorsal surface. The differences in the response of the various wing surfaces to cautery may simply be due to the fact that their development is asynchronous so that at any one time they are not in equivalent stages of colour pattern determination. On this assumption, the forewing would be expected to pass through a stage in which eyespots could be induced by cautery some time prior to pupation. On the other hand, the various wing surfaces may simply differ in their threshold to the ‘cautery-response’, so that cautery only evokes a morphological effect in regions where this threshold is quite low. This latter view seems at this time to be the most reasonable, as will be discussed below.

Coldshock-induced pattern modifications in Precis were described by Nijhout (1984a), who showed that, for a given wing-cell, the modified patterns could be arranged in a smooth linear morphocline from nearly normal to highly aberrant. With the exception of the dorsal hindwing, progressively more affected patterns in a morphocline always exhibited a progressive diminution in the diameter of eyespots.

The large eyespot on the dorsal hindwing of Precis, however, undergoes a most unusual modification of shape that has no counterpart on any of the other wing surfaces, nor have such modifications been seen in any other species I have studied so far. Figure 3 illustrates several of these patterns. In the least affected individuals the outer light ring of the eyespot becomes slightly broader than usual (Fig. 3A). This effect seems to come about through a diminution in the diameter of the central dark disk of the eyespot though this cannot be tested objectively because there is natural individual variability in the diameter of the eyespot, and both wings of an individual are always affected identically. In slightly more severe pattern modifications the dark and light outer rings of the eyespot become extended proximally or distally, or in both directions, depending on the individual (Fig. 3B–D). More dramatic aberrations also affect the central disc of the eyespot, which tends to become extended proximally, and at least part of the outline of the eyespot becomes smeared out and indistinct. In the most extreme cases the entire anterior half of the hindwing is set off from the posterior half by a broad black streak, homologous to the black outer ring of the eyespot (Fig. 3E, F).

In contrast to the coldshock-induced pattern modifications of the other wing surfaces (Nijhout, 1984a), those on the dorsal hindwing could not be arranged into a smooth morphological series. As a general trend, there is a progressive elongation of the anterior eyespot along the proximodistal wing axis (evident in Fig. 3A→E), but out of some 40 moderately to severely affected individuals, no two resembled each other either in detail or in overall shape of the pattern modification, which suggests the operation of a pattern generating mechanism with a relatively large amount of noise. The aberrant patterns on the left and right wings of a given individual, however, were always identical or nearly so.

The simplest interpretation of the findings presented above and illustrated in Fig. 1 is that cautery somehow mimicks a focus and causes the local synthesis of a morphogen identical to, or equivalent to, the one normally produced by a focus and responsible for the formation of native eyespots. This interpretation is, however, inconsistent with all earlier known effects of cautery on colour pattern formation. Cautery, when placed on presumptive source cells in the dorsal forewing of Precis, clearly eliminates those sources and abolishes development of an eyespot (Nijhout, 1980a). Cauteries placed at other sites of the dorsal forewing of Precis have no effect on the colour pattern (Nijhout, 1980a). In Ephestia cauteries cause local distortions of the pattern which have been interpreted as being due to the fact that cauterized cells form a local obstacle to the propagation of a pattern-determining signal or morphogen (Kühn & Von Engelhardt, 1933; Murray, 1981). While it is possible that these multifarious effects of cautery (elimination of a source [Precis forewing], apparent induction of a source [Precis hindwing], apparent obstacle to signal propagation [Ephestia forewing]), are due to fundamental species- and wing disk-specific differences in the mechanism of pattern development, it is more likely that these various effects are simply different manifestations of a common response to cautery. It is, in fact, possible to construct a single model that can account for all the known effects of cautery.

If it is assumed that colour patterns are determined at least in part by chemical morphogens that emanate from discrete sources (Nijhout, 1978, 1980a,b, 1984b), and if the hypothesis that cautery induces a local source of morphogen is rejected as inconsistent with other findings, then four other potential effects of cautery at the cell and tissue level are plausible and may be considered in formulating a hypothesis. 1) Cautery could cause the local destruction of a morphogen as an incidental response to wounding or wound healing. The site of cautery would then behave as a sink for morphogen. 2) Cautery could locally alter the sensitivity of cells to morphogen. Such a local raising or lowering of a threshold could lead to distortions of the colour pattern. 3) Cautery could cause a local delay in development and determination while would healing takes place. Finally, 4) cautery could simply kill cells, making them incapable of producing, transmitting or destroying morphogen. This is the hypothesis Kühn & Von Engelhardt (1933) used to explain the effects of cautery on pattern modification in Ephestia.

The first two hypotheses may be different aspects of the same mechanism. In general, the threshold of sensitivity of cells to a given substance will be set by the concentration (and binding affinity) of other chemicals in the cell with which the first reacts. The distribution of such a threshold-determining substance represents what I have called the interpretation landscape (Nijhout, 1978,1984b). In order to produce a circle from a point source, it is necessary that the interpretation landscape be flat, which is to say, that the substance with which the local morphogen interacts be homogeneously distributed at some finite concentration.

The cautery-induced eyespots on the hindwing (Fig. 1) and the distortions in the shape of the forewing eyespot (Fig. 2) can be readily explained by assuming that cautery causes a local depletion of the substance that makes up the interpretation landscape. Figure 4 shows the result of a computer simulation based on this hypothesis which mimicks the morphological effects of cautery reasonably well.

The third hypothesis, that cautery and wound healing cause a local delay in determination is not, in principle, unreasonable. Such a developmental delay could result in the formation of an eyespot if it occurred in the presence of a continually rising level of morphogen, and if the delay was somehow graded from the site of cautery outwards. Morphological evidence, however, does not provide evidence for such a delay. Scale development, which begins immediately after the period for colour pattern determination (Nijhout, 1980b), is synchronous at the site of cautery with scale development elsehwere on the wing, as is the initiation of pigment synthesis. Furthermore, if such a model is also to be used to explain the effect of cautery on the colour pattern of Ephestia (see below), many of the assumptions must be reversed (e.g. either the cautery must accelerate determination, or the sources of Kühn & Von Engelhardt must be sinks).

The differences in the observed responses to cautery and coldshock on forewing and hindwing are unlikely to be due to differences in the timing of equivalent developmental stages on each wing surface. Other developmental events such as mitosis, which coincides with the period of colour pattern determination, and scale differentiation and pigment synthesis, which follow shortly thereafter (Nijhout, 1980b), are synchronous on both wing surfaces and there is no a priori reason for believing that colour pattern determination would be uniquely asynchronous. Furthermore, I have done cauteries of the dorsal forewing over a broad time span, from the moment of pupation to 36 h after the termination of colour pattern determination (Nijhout, 1980a), and I have recently done surgery on imaginai disks in situ at times from 12 h to 4 days prior to pupation, and at no time have I observed effects that even remotely resemble the induced eyespots on the dorsal hindwing. Since the hindwings are susceptible to eyespot induction for at least a 24 h period it should have been possible to detect an equivalent period for the forewing, if it existed. Likewise, coldshocks have been done on animals from 24h prior to pupation to 48 h after pupation. Both forewing and hindwing were found to have the same sensitive period for the induction of colour pattern aberrations (0 to 24 h after pupation), and in hundreds of coldshocked animals I have never found morphological aberrations of the forewing eyespot that resemble those of the eyespot on the hindwing (Nijhout, 1984a). Differences in the responses of forewing and hindwing to cautery and temperature shock are thus probably not simply due to asynchronies in pattern development on these two surfaces. On the model presented above one would have to assume that the effectiveness of the cautery-induced sink differed on fore and hindwing.

Cauteries on the wings of Ephestia have an effect on the colour pattern that appears to be the exact opposite of what I have described above for Precis. Whereas in Precis the site of cautery becomes enclosed by a ring of pigment, in Ephestia the course of pigment bands becomes altered so as to exclude the site of cautery from the central pattern field (Fig. 5A; see Kühn & Von Engelhardt, 1933, for numerous illustrations). The pattern alterations in Ephestia have traditionally been ascribed to the fact that cautery sets up a barrier to the diffusion of morphogen (Kühn & Von Engelhardt, 1933; Murray, 1981).

I have simulated the effect of imposing a barrier to diffusion on the morphology of the pigment bands in Ephestia. These simulations revealed that while the gross features of the experimental results illustrated by Kühn and Von Engelhardt (deviation of bands around the site of cautery) are readily mimicked, there are two morphological details that proved difficult or impossible to simulate, namely, 1) the development of a border of light-coloured scales (identical in colour to the pigment bands) around the site of cautery, and 2) the frequent occurrence of a narrow bridge of light-coloured scales that connects the border around the site of cautery with the nearest pigment band (Fig. 5A). A ring of pigment that runs completely around the site of cautery proved impossible to obtain (Fig. 5B), due to the fact that diffusing morphogen tends to bank up slightly on the proximal (upstream) side of the barrier. By contrast, when a cautery was simulated as a sink for morphogen, the bordering ring of pigment around the cautery site was almost always obtained, and a bridge between it and one of the simulated pigment bands could easily be made to appear by an appropriate choice of thresholds (Fig. 5C, D). It appears that the effect of cautery on the colour pattern of Ephestia can be explained better by assuming that cautery induces a local sink for morphogen than by assuming that it merely presents a barrier to its diffusion.

In Precis, cautery of the hindwing usually induces the outer (black and yellow) rings of the eyespot, and very light cautery induces only a local spot of black pigment (Fig. 1A), homologous to the black outer ring of the eyespot. Thus fairly small disturbances in the background field are evidently sufficient to establish the conditions under which the outer rings of the eyespot are normally induced.

This observation may help explain the erratic development of the outer rings of the large eyespot after coldshock (Fig. 3). I have previously shown that the development of colour pattern aberrations (phenocopies) after coldshock is due to alteration of a single developmental parameter (Nijhout, 1984a), and that circumstantial evidence strongly suggests that this parameter is associated with the formation of the interpretation landscape (Nijhout, 1984b). If coldshock somehow lowers the level of the interpretation landscape, and if, as the cautery experiments suggest, the level of the interpretation landscape in a normal wing is already close to critical for the determination of the outer bands of an eyespot, then it is possible that after coldshock the interpretation landscape could fall to a critical level for pattern induction over wide regions of the wing. The irregular character of coldshock-induced patterns could then be due to irregularities in the normal distribution of the putative interpretation landscape-morphogen. This interpretation finds circumstantial support in the observation that the size and shape of the eyespots on the dorsal hindwing of Precis exhibit an uncommon degree of individual variability (illustrations in Comstock, 1927).

Figure 3 shows that a sector of the wing that contains the smaller of the two eyespots remains relatively unaffected by coldshock, while cauteries in this area induce eyespots as effectively as they do elsewhere on the wing (Fig. 1). It is not clear how this relative stability to coldshock is to be explained. I am unable to detect a developmental asynchrony between these two regions of the wing, but it is noteworthy that the demarcation between the affected and unaffected areas, illustrated in Fig. 3E,F roughly corresponds in position to a compartment boundary detected by Sibatani (1982) in homeotic transformations of the wings of many species of Lepidoptera.

I wish to thank Mary Nijhout and Greg Wray for critical comments on the manuscript, Laura Williams for drawing Figure 5A and for technical assistance with the experimental work, and Greg Wray for help in writing the interactive features of the computer programs used in this study. This work was supported in part by the Duke University Research Council and by Grant PCM-8214535 from the National Science Foundation.