The eye is hailed as a paragon of organismal complexity, an organ of sophisticated design and many interacting parts
. Here we report how an insect’s eye morphology and physiology can be regulated by developmental rearing temperature in a fashion that is likely to be adaptive to adults emerging in alternating seasons.
Our work documents sexual dimorphism and plasticity in opsin levels in B. anynana in a complex way, cued by developmental rearing temperature. At low rearing temperature, males display significantly higher levels of Blue and LW opsin mRNA levels than females, and females reared at low temperature display significantly lower UV, Blue, and LW opsin mRNA levels than females reared at high temperature. We demonstrate, for the first time, that developmental rearing temperature induces changes in opsin expression levels in an insect’s eye. However, plasticity in type or level of opsin expression was previously documented in a variety of species responding to different environmental conditions: Opsin levels change with light rearing environment in African cichlids
, with the diurnal cycle in Limulus, and with age in Drosophila; whereas new opsins are induced upon maturation in the European eel
, or upon a change in lifestyle in salmon
. In addition, sexual dimorphism in opsin spatial expression patterns
 and in the presence/absence of a non-opsin filter pigment
 were previously documented in butterflies.
According to our original hypothesis, and barring any developmental constraints, we expected B. anynana to have evolved plasticity in its visual system, due to physiological costs associated with vision
. In particular, we expected the visual system to decrease in capacity in non-choosy courters. We found that physiological and morphological changes conformed, in part, to our predictions.
The non-choosy DS females displayed lower levels of UV, blue, and LW opsin transcript than the choosy WS females. Assuming that opsin mRNA is being actively translated to protein, DS females appear to have reduced the costs associated with photoreceptor energy consumption
 in exchange for reduced visual function. To measure the impact of opsin levels on visual function, in vivo intracellular microelectrode recordings would provide direct physiological information on the costs and benefits of opsin production in Bicyclus. The DS females, who mate indiscriminately with males with or without the dorsal eyespot ornaments
, can afford such loss of visual function. Males, however, including the equally indiscriminate WS males, maintained high levels of opsin expression across seasons and did not conform to our original predictions.
Our results for eye size and facet size plasticity also only partly support our original hypothesis. We predicted that choosy individuals should develop either higher acuity or greater sensitivity to light to evaluate the small dorsal eyespot centers, i.e. WS females and DS males should have more facets and/or larger facets than their non-choosy DS and WS same-sex forms. We found that facet size was especially large in the choosy DS males relative to WS males, and choosy WS females had more facets than non-choosy DS females, but the reverse prediction did not pan out for female facet size or male facet number. Facet number and facet size were always larger in males relative to females of both forms, and facet number was always larger in WS versus DS individuals, with no significant sex by season interaction. In order to explain these results, we sought alternative explanations for the observed patterns of sexual dimorphism and plasticity that move away from examining B. anynana eyes as useful only for evaluating sexual ornaments in the context of mate choice.
Smaller facets at warmer WS temperatures may be explained either by a biophysical constraint, the “temperature-size rule”
, or by natural selection for improved light sensitivity in the DS or visual acuity in the WS, for both sexes. The “temperature size rule” states that the rate of cell division increases more than the rate of cell growth with increasing temperature, and helps explains the pervasive pattern of small bodies at high temperatures (but see
 for a second explanation of this latter phenomenon). Animals experiencing high temperature during development, thus, will end up reaching maturity with the same number of cells but with smaller cells
. Natural selection-driven alternatives to this biophysical constraint, could be that larger facets, often associated with activity at lower light levels
, are selected for in the DS, but currently we have no indication that light levels differ between the dry and wet seasons in Malawi, or that the two seasonal forms are active at different times of the day. Another natural selection-driven alternative is that smaller facets in the WS are actually adaptive as this leads to lower inter-ommatidial angles and improves visual acuity
 in the WS.
Regardless of the forces driving smaller facet size in WS eyes, our study suggests that visual demands are lower in the DS because DS eyes are 13% smaller than WS eyes across sexes. Lower visual demands may be related to lower environmental complexity in the DS or with lower activity levels in DS butterflies. Lower temperature is known to decrease butterfly activity levels in the field
[43, 44], and lab observations of B. anynana have indicated that DS butterflies are generally less active than their WS counterparts. Larger eyes in the WS, on the other hand, would allow the more active, and thus more conspicuous, WS butterflies to search for mates and thwart predators more efficiently. Alternatively, visual demands in the DS are trading off against other energetically expensive demands, such as survival through the long dry season
Sex-specific butterfly behaviors, beyond mate ornament discrimination, may also contribute to explain why the larger-bodied females have fewer facets than the smaller-bodied males, why opsin expression levels were high in males of both seasonal forms, and why DS females have the lower opsin levels of all four groups. Males of B. anynana had 28% larger eyes and 12% larger facets than the larger females across temperatures. This is a different pattern from that found in Drosophila, where the larger females also have larger eyes
[45, 46]. Our data, however, matches previous studies of eye size sexual dimorphism in other butterfly species
[47–49], and, although developmental constraints cannot be ruled out, the dimorphism is likely to be the result of male-limited activities such as mate searching, and territory defense typical of satyrid butterflies
[44, 50]. These activities, as well as their side effects of becoming more visible to predators, may require males to maintain large eyes and high opsin expression across seasons. B. anynana males have a typical perch-and-chase strategy
, and males with better vision are expected to have an advantage at detecting passing females, competing males, or nearby predators. The plastic courtship roles may take over only once males have localized a female
. On the other hand, female-limited searches for oviposition sites, which may be suspended in the DS (due to ovary dormancy
), may lead, through relaxed selection, to reductions in eye size and opsin levels in DS females due to their high maintenance costs
Plasticity of eye size in B. anynana appears to be operating through the control of resource allocation between different body parts. Resource competition between imaginal discs in holometabolous insects reared at a constant temperature can give rise to size trade-offs in adult body parts, such as eyes and wings. This competition takes place during the pre-pupal and pupal stages because growth of the imaginal discs happens in a closed system once the larva has stopped feeding
[52–55]. It appears that in B. anynana, high rearing temperature is cueing development to shift resource allocation away from wings and into eyes (Figure
3). Additional experimentation will be required to test whether these plastic patterns of allocation from wings to eyes are adaptive.