External morphology of eyes and Nebenaugen of caridean decapods–ecological and systematic considerations

Most caridean decapods have compound eyes of the reflecting superposition kind, and additionally some possess an accessory eye-like organ of unknown function, also referred to as the nebenauge. We examined 308 caridean genera to assess the general morphology of the eye, rostrum length, eye diameter and the presence or absence and, when present, the diameter of the nebenauge. We have attempted to relate these data to ecological and taxonomic considerations. We consider there to be 6 distinct eye types based on the margin between the eyestalk and cornea. The presence of nebenaugen appears to be generally linked to an active lifestyle, as evidenced by the fact that species that have nebenaugen tend to have larger eyes and are more likely to have a distinct rostrum. We suggest that the inconsistencies in its presence/absence under both systematic and ecological lenses may indicate that when present it has various roles relating to behavioural and physiological rhythms.

113 The function and systematic relevance of the nebenauge of decapods remains unclear (Gaten et al 1992, 114 Ugolini & Borgioli, 1993) despite a long history of interest in the structure of the main eye and the 115 physiological importance of the closely associated sinus gland complex. Also there have been no 116 extensive examinations of the variation in gross eye design amongst caridean decapods and few 117 decapod systematicists recognise its potential as a useful diagnostic character.
118 Aims 119 The aims of this investigation are to examine the potential taxonomic significance and likely ecological 120 role of the nebenauge of caridean shrimps through an extensive examination of its presence absence 121 and relative size in representatives of every genera of caridean. We also assess the variation in gross 122 eye structure of decapods generally and in relation to the presence/absence of the nebenauge. 134 Although troglobitic genera were examined we found they do not have well developed eyes and so they 135 were excluded from our analysis. Where there were several genera to choose from in the collection we 136 examined the largest representative or used a species where there were numerous individuals available.
137 Details of a further 133 genera were extracted from the literature. For the most part the relevant 138 taxonomic papers were found after consultation with De Grave & Fransen (2011). When possible, 139 carapace length, rostrum length, nebenauge presence were noted, either by direct reference to the text 140 or through observations taken from the drawings. Preliminary analysis of the dataset (see Supplemental 141 Information) indicated that measurements taken from papers were not robust enough to be anything 142 more than indicative and so were only used for presence absence of nebenaugen and rostrum length 143 classification. Analysis of a dataset with date since collection of samples indicated no effect of sample 144 age on superficial eye morphology (see Supplemental information).
145 Apportioning precise depth ranges on the basis of a few samples would not be wise so animals were 146 classified according to their deepest recorded depth distribution into freshwater, coastal (0-99 m), shelf 147 (100-300 m) and deep (>300 m) based on available information from the literature. Freshwater 148 specimens were not classified by depth because of the wider range of optical properties of lentic and 149 lotic environments. Animals were also categorised by habitat (temperate, tropical, deepsea) and as 150 either commensal or free living.

Analysis
152 All analyses were carried out using the statistical package R (Ihaka & Gentleman, 1996 157 We found that eyes could generally be classed into 6 major types according to the shape of the margin 158 between the eyestalk and the cornea (Figure 2), general shape and further classified by the presence or 159 absence of the nebenauge.
160 The distribution of eye types among families was not random (X 2 =285.5, d.f.=11,308, p<0.0001) and 161 neither was the distribution of nebenaugen among them (See Table 1). Although present in 14 families, 162 the most basic eye design (Type 1) which consists of a tube shaped eyestalk and a simple hemispherical 163 eye was the most common because it was favoured by the most speciose sub-family, the Pontoniinae.
164 Fifteen families possessed the next most common, Type 2 eyes which is the only classification that has 165 more species with nebenaugen than without. Type 6, which generally consisted of reduced eyes, was 166 the only classification where none of the examples examined possessed a nebenauge.

Nebenaugen
168 We examined the eyes of 308 non-troglobitic genera from 26 families ( Table 1) and found that 88 169 possessed a discernible nebenauge (28.6%). We found that the distribution of nebenaugen across 170 caridean families, where we sampled more than the minimum of 5 genera, deviates from a 50:50 171 distribution overall (p<0.001, Table 1). There appear to be some links between its presence/absence and 172 systematics with non-random differences in the prevalence of nebenaugen. The distribution of 173 nebenaugen amongst genera by family could be split into those where genera had about a 25-50% 174 chance of having one (Hippolytidae, Oplophoridae and Pandalidae), those where they occurred 175 infrequently (Pontoniinae and Atyidae) and those where none were found for any species (Alpheidae 176 and Crangonidae). The only family where significantly more genera were found to have nebenaugen 177 than not were the Palaemonidae. There are no families where all genera possess a nebenauge. Analysis 223 Although the link may be considered tenuous given the numbers of pelagic decapods that have small 224 rostra (Sarda, Company & Costa, 2005), we have used the presence of a large rostrum, which would be a 225 physical hindrance to an endocommensal or infaunal habit, as an indicator that these species are likely 226 to be free living and active swimmers. Our data suggest that animals that have a long rostrum relative 227 to carapace length tend to have larger eyes and are more likely to have nebenaugen (Figures 3 & 6). 236 In marine species the likely relationships between habitat, lifestyle, gross eye morphology and eye size 237 will be further complicated by consideration of habitat depth. Depth has been shown to have a strong 238 impact on the morphology and function of decapod eyes . Animals found in 239 shallower water tend to have "faster" eyes with a greater capacity for light adaptation than those from 290 Conclusion 291 We conclude that, although as with previous investigations there appears to be no clear taxonomic 292 pattern in its distribution by family and although we cannot ascribe a function to the nebenauge of 293 caridean decapods, its presence appears to be linked to lifestyle. It is more frequently present as eye 294 size of a species increases and shows an association with relative rostrum length. If size is an indication 295 of ecological importance then the fact that it is larger in species found below 100 m suggests that it has 296 a particularly important role in some deep-water species. With regard to the superficial morphology of 297 caridean eyes, we have found a log-log relationship between carapace length and eye diameter and 298 suggest that there are 6 identifiable eye morphotypes based on general morphology and the margin 299 between eyestalk and cornea.