Electronic Supplementary Results

Electronic supplementary results

Here we present additional analysis on the population structure (body size, life-history stage and gender) of the obligate coral-dwellers when considering the Pontoniinae shrimp and Trapezia separately:

Body size

The Pontoniinae coral-dwellers obligate model showed no significant relationship with coral colony size or biogeography. Brothers and Egmont had the highest mean Pontoniinae body size and Diego Garcia showed the lowest (Fig. S.2.a). At the individual atoll scale, there was a significant interaction with coral colony size at Egmont (t =2.45, p =0.02), indicating that at Egmont coral colony size did affect Pontoniinae shrimp obligate body size. Unlike the total coral obligate model, the shrimp obligate model showed under-dispersion (dispersion parameter =0.33) demonstrating the small variation in body size among Pontoniinaeshrimp species maybe driven by ecological processes such as competition. In the Trapezia obligate assemblage the largest mean body size occurred at Egmont (Fig. S.2.a), which most likely resulted from the presence of two adult Trapezia in two different coral colonies at Egmont. However, the model shows no significant relationship with biogeography or coral colony size (Fig. S.2.a & b). In contrast to the shrimp, the Trapezia obligate model is highly over-dispersed (dispersion parameter =2.83) representing the variation in body size between individual Trapezia.

Life-history stage & gender

Within the Trapezia obligate assemblage the lowest difference in the proportion of adults to juveniles occurred at Egmont, whilst no adult Trapezia were found at all at Eagle (Fig. S.3.a). However, the Trapezia obligate assemblage model showed no significant difference in the proportion of the two life-history stages across locations or with coral colony size (Fig. S.3.a & b). Within the Pontoniinae shrimp obligate assemblage no juveniles were found at Brothers, Eagle, or Peros Banhos, and only one or two individuals at the other atolls and islands (Fig. S3.a). The Pontoniinae shrimp life-stage model showed a significant interaction affect of coral colony size and biogeography (F =13.24, p <0.01) on the proportion of adults and juveniles in the assemblage. To investigate the affect of no Pontoniinae shrimp juveniles found at Brothers, Eagle, or Peros Banhos, the Pontoniinaeshrimp life-history stage counts from these three locations were grouped together. This model showed the same significant interaction affect of coral colony size and biogeography (F =22.91, p <0.01). The shrimp obligate model showed under-dispersion (dispersion parameter =0.15), demonstrating the high ratio of adults to juveniles.

Within the Pontoniinaeshrimp obligate assemblage the proportion of females to males was highest at Eagle, whilst at Salomon and Diego Garcia the proportion of females to males was relatively even (Fig. S.4.a). Neither coral colony size nor biogeography were significant drivers of shrimp gender at the archipelago assemblage level. The Trapezia obligate assemblage model failed to converge, probably because of the small number of adults, only 13, in the assemblage. However Fig. S.4.a illustrates the higher proportion of males to females across all the atolls and islands, with the exception of Diego Garcia where there were 2 female and 1 male Trapezia (Table 2). There was also no obvious trend with coral colony size (Fig. S.4.b).