Caridina typus, AND C. AFRICANA

COMPARISON OF C. TYPUS AND C. AFRICANA View in CoL

Both Caridina species exhibit radically different levels of phylogeographical structuring when compared across the same geographical scale, as has been demonstrated for other congeneric species at varying spatial scales. Page & Hughes (2007) recovered large disparity amongst four cryptic species of eastern Australian populations from the Caridina indistincta View in CoL complex in their scales of intraspecific divergence, geographical distribution and population structure. Likewise, in southwestern Japan, Fujita et al. (2016) demonstrated that Caridina leucosticta Stimpson, 1860 View in CoL exhibited a genetically heterogeneous population structure compared with the homogeneous pattern seen in C. typus and View in CoL Caridina multidentata Stimpson, 1860 View in CoL .

One explanation for the disparity in phylogeographical structure in sympatric, congeneric shrimp taxa is life-history variation, in particular larval dispersal ability. Amphidromous atyid and palaemonid shrimps have high fecundity and relatively small eggs at maturity, with the inferred lengthened larval phase facilitating dispersal at a range of geographical scales ( Bauer, 2013; Wood et al., 2017). Indeed, it has been speculated that the extent of larval movement can be attributed to species-specific larval salinity tolerances and the number of larval stages ( Fujita et al., 2016), with taxa that have short larval stages and lower salinity tolerances having a lower dispersal ability. This results in genetic heterogeneity and smaller geographical ranges. In contrast, species that spend a longer time in marine plankton as larvae show the opposite [ Fujita et al., 2016; but see Santos (2006) and Craft et al. (2008) for exceptions]. The present analysis indicates considerable genetic homogeneity and higher levels of gene flow across South African populations of C. typus View in CoL , with individuals belonging to the ARC cryptic lineage ( Bernardes et al., 2017); uncorrected pairwise COI divergence of between 1% ( Mauritius) and 7.5% ( New Caledonia). The ARC lineage occurs across the margins and islands of the Indian Ocean, to the Philippines and as far east as Vanuatu and northwards to Taiwan and Okinawa and does exhibit geographical heterogeneity on that wider scale ( Bernardes et al., 2017). Fujita et al. (2016) previously indicated that dispersal ability in the ARC cryptic lineage may be linked with life-history traits, with the Japanese population of C. typus View in CoL having a relatively high number (nine) of zoeal stages, with a high optimal salinity of 25.5 ppm for the first zoeal stages ( Nakahara et al., 2005, 2007), facilitating movement between hydrological systems.

In contrast, the high level of COI genetic differentiation observed for C. africana is indicative of restricted gene flow across drainage basins. Although particular larval traits of C. africana remain undocumented, given that a number of localities from which this species was collected are land locked ( Table 1; Fig. 4 View Figure 4 ), it is clear that reproductive plasticity occurs in this species. Riverine populations may have retained an amphidromous life-history strategy, perhaps with a lower number of zoeal stages and/or lower salinity tolerance. Land-locked populations, in contrast, must have adapted this reproductive strategy to complete their life cycle in freshwater. Intraspecific variable reproductive strategies in relationship to the distance away from the coast or to the population being land locked have been documented in a number of freshwater Palaemonidae . For example, in the widely distributed South American species Macrobrachium amazonicum (Heller, 1862) , geographically delineated variation in life-history traits, from coastal amphidromous populations to far-inland populations, has been demonstrated ( Vergamini et al., 2011; Hayd & Anger, 2013). However, such intraspecific variation in life-history strategy has not yet been reliably documented within Atyidae .