Variation in Numbers of Scleral Ossicles and Their Phylogenetic Transformations within the Pelecaniformes

•We examined scierai rings from 44 species of Pelecaniformes and found nonrandom variation in numbers of scierai ossicles among genera, but little or no variability within genera. Phaethon, Fregata, and Pelecanus retain the primitive 15 ossicles per ring, while the most recent common ancestor of the Sulae (Phalacrocoracidae, Anhinga, and Sulidae) is inferred to have had a derived reduction to 12 or 13 ossicles. Within the Sulidae, Sula {sensu stricto) exhibits further reduction to 10 ossicles. These patterns of ossicle reduction are congruent with both Cracraft's hypothesis of pelecaniform relationships (1985) and that of Sibley et al. (1988). The presence of scierai rings in museum specimens is significantly greater for Phaethon and Fregata, and less for Pelecanus, than would be expected from a random distribution. We conclude that the scierai ring is of potential systematic importance, and we make recommendations for its preservation in museum collections. Received 11 August 1988, accepted 25 January 1989. THE scierai ring (annulus ossicularis sclerae) is a ring of smail overlapping piatelilce bones, ttie scierai ossicles (ossicula sclerae), found w^ittiin ttie sciera in ttie corneal liemispiiere of tiie eye between the retinal margin and tbie conjunctival ring (Edinger 1929, Martin 1985). Th.e function of the scierai ring as a v^fhole, and the individual ossicles in particular, is poorly known. Two functions are most often proposed for the scierai ring: first, as support and protection of the eye in tiie region where it is found; alternatively, as an attachment for the ciliary muscles, specifically, the anterior corneal muscle {m. cornealis anterior), which suggests a role in corneal accommodation (Lemmrich 1931, King and McLelland 1984, Martin 1985). Curtis and Miller (1938) also suggested that eccentricity of the ring aperture may facilitate binocular vision. The arrangement of the ossicles themselves is even more problematic, although Lemmrich (1931) proposed that having a scierai ring composed of overlapping plates permits growth of the ring during ontogeny. Lemmrich also proposed that the differences in ossicle number among taxa may be attributed to "growth organization" (Organisationmerkmal) particular to taxa. Because the scierai ring exists in a wide variety of vertebrates (Edinger 1929) including nonavian dinosaurs (references in de Queiroz and Good 1988), its presence in birds appears to be a retained primitive character. Therefore, any explanation of the ring's function within birds must also consider its widespread occurrence in Vertebrata. Whatever function the scierai ring may serve, the number of ossicles per ring varies in a nonrandom pattern among some avian taxa (Lemmrich 1931, Curtis and Miller 1938, de Queiroz and Good 1988). The use of the scierai ring as a systematic character in birds has been limited. Studies have been descriptive surveys, conducted primarily at the ordinal or familial levels, on ring morphology, function, and variation in ossicle number and position (e.g. Lemmrich 1931, Curtis and Miller 1938). While conducting two independent research projects, one on the scierai ossicles and phylogenetic relationships of the Hoatzin (de Queiroz and Good 1988), and the other on the phylogenetic relationships within the Sulidae (Olson and Warheit 1988, Warheit in prep.), we discovered what appears to be phylogenetically informative patterns of variation in numbers of scierai ossicles among the pelecaniform genera. We summarized the numbers of scierai ossicles for extant species of Pelecaniformes, and examined these patterns in the context of current hypotheses about pelecaniform relationships. As a result of our survey, we also found that scierai rings are not satisfactorily main383 The Auk 106: 383-388. July 1989 384 WARHEIT, GOOD, AND DE QUEIROZ [Auk, Vol. 106 tained in osteological collections, and are absent even in some specimens that are considered "complete." Accordingly, we make recommendations for the preparation of scierai rings in osteological specimens. MATERIALS AND METHODS We sampled 695 skeletons from 51 species of pelecaniforms. The number of specimens sampled for a particular species was determined by its availability in the collections. For species with a limited number of skeletons, all specimens were examined. If the sample size for a species was large (« > 20), skeletons were chosen randomly, but with an effort to maintain equal numbers of males and females. We recorded the number of complete rings present (0, 1, or 2) for each specimen. When ossicle overlap is extreme, some plates may be visible on only one side of the ring. Therefore, the number of ossicles that constitute a ring was determined by counting plates on both the inner and outer sides. For specimens with two rings, both rings were examined. In most cases (87%) the number of ossicles in the left and right rings was the same, but occasionally the number differed (see Curtis and Miller 1938 for univariate statistics and discussion of constancy of scierai ossicle numbers between right and left eyes for various taxa of birds). Mean counts were recorded for those specimens with both left and right rings preserved. For example, if the number of ossicles in the left ring was 14 and the right ring was 15, a count of 14.5 was used for that specimen. This was done because using the total number of scierai rings to calculate the mode, rather than the preferred total number of specimens, would not only violate the assumption of independent samples, but would also bias the mode towards specimens with two rings over those with only one ring. We summarized the data in two ways. First, modal numbers of ossicles are reported for 44 of the 51 species surveyed (no scierai rings were available for 7 species). Second, the availability of ossicles for each species was determined by comparing the number of specimens containing ossicles (either one or two) to the total number of specimens examined. Specimens with broken or disarticulated rings were excluded.