Variation in orchid seed size and shape can be linked to phylogenetic relationships, habitat preferences, germination behaviour or dispersal strategies. To investigate this, we compared 45 orchid species from 29 genera collected across different localities in Cameroon using optical microscopy. We categorized each species according to lifeform (38 epiphytic vs. 7 terrestrial), altitudinal range (11 mountain vs. 34 lowland) and geographic distribution (28 widespread vs. 17 range-restricted). We analysed seed morphology using phylogenetic signal tests, analysis of variance and principal component analysis. Our results confirm a clear distinction between epiphytic and terrestrial species, with intermediate morphologies observed in genera encompassing species with both lifeforms (Cynorkis, Graphorkis, Habenaria and Liparis). Certain traits, such as seed length and seed air space, show a strong phylogenetic signal, suggesting that these traits are more linked to ancient evolutionary history than to recent ecological adaptation. Among the 38 epiphytic species, no consistent relations were found between seed traits and either geographic range or altitudinal distribution. Our findings suggest that the variation observed in seed morphology among African orchids is influenced more by phylogenetic relationships than by present-day distribution.

It appears that squid statoliths cannot yet be regarded as accurate an ageing tool as fish otoliths. Statoliths from the same pair, prepared differently for viewing and counting increments, were compared. Increment counts do not imply age in days, because this was not validated. One statolith from each pair was examined by light microscopy (LM) after preparation following a new method. The other was viewed by Scanning Electron microscopy (SEM) with a modified etching solution. Shape of each statolith was similar when compared by multiple regression analysis (11 variables, n=53). There was a weak but significant difference between sexes (statoliths of females were slightly larger). All other differences were insignificant. Microscopic observation and increment counts of increments were successfully carried out for 37 pairs of statoliths. Significant differences between two independent counts were found for the LM method, but no significant differences were found between two independent SEM counts. Counts were significantly different when interpreted by both LM and SEM, probably because of poor resolution in the LM readings and over-resolution (growth layers prominent and numerous) in those read by SEM. Recommendations are made on how ageing studies, based on statoliths, should be structured and the results evaluated.