JOURNAL OF MORPHOLOGY 221:261-276 (1994) Com parative Morphology of Caecilian Sperm (Amp h i bi a: Gym nop h ion a) MAFWALEE H. WAKE Department of Integrative Biology and Museum of Vertebrate Zoology, University of California, Berkeley, California 94720 ABSTRACT The morphology of mature sperm from the testes of 22 genera and 29 species representing all five families of caecilians (Amphibia: Gymnoph- iona) was examined at the light microscope level in order to: (1)determine the effectiveness of silver-staining techniques on long-preserved, rare material, (2) assess the comparative morphology of sperm quantitatively, (3) compare pat- terns of caecilian sperm morphology with that of other amphibians, and (4) determine if sperm morphology presents any characters useful for systematic analysis. Although patterns of sperm morphology are quite consistent intrage- nerically and intrafamilially, there are inconsistencies as well. Two major types of sperm occur among caecilians: those with very long heads and pointed acrosomes, and those with shorter, wider heads and blunt acrosomes. Several taxa have sperm with undulating membranes on the flagella, but limitations of the technique likely prevented full determination of tail morphology among all taxa. Cluster analysis is more appropriate for these data than is phylogenetic analysis. cc: 1994 Wiley-Liss, Inc. Examination of sperm for purposes of describ- ('70), in a general discussion of aspects of ing comparative sperm morphology within sperm morphology, and especially Fouquette and across lineages and for systematic assess- and Delahoussaye ('77, '921, Morrisett ('741, ment is enjoying a renaissance. The publica- and Wortham et al., ('77, '821, and recently tion of Jamieson's volume ('91) on compara- Lee and Jamieson ('92a,b) and Jamieson et tive morphology and systematics of fish al., ('93) have considered the phylogenetic sperm, with a discussion of techniques and a significance of sperm morphology in amphib- general assessment of the utility of sperm ians, particularly of families of frogs and ultrastructural characters for systematic salamanders. studies of both invertebrates and verte- Descriptions of caecilian sperm morphol- brates, presents an extreme, but useful, eluci- ogy are limited to analyses of patterns of dation of the potential value of sperm charac- spermatogenesis in individual species (de Sa ters to phylogenetic reconstruction (see and Berois, '86; Chthonerpeton indistinc- Discussion). tum; Exbrayat and Sentis, '82; Exbrayat, '86: Among amphibians, sperm have been de- Typhlonectes compressicaudus; Seshachar, scribed for several taxa of frogs and sala- '36, '37a,b: Ichthyophis glutinosus, '39: manders, usually in the context of assessing Uraeotyphlus narayani, '40: several species, patterns of spermatogenesis (e.g., Burgos and '42a,b: Siphonops annulatus and Schistome- Fawcett, '56; Baker, '62, '63, '66; Austin and topum [then Dermophis] gregorii, '43: Ich- Baker, '64; Barker and Biesele, '67; Brandon thyophis glutinosus, '45: Uraeotyphlus naray- et al., '74) or describing sperm at the light ani; Wake, '68; Gymnopis multiplicata, with microscope or ultrastructural levels (e.g., notes on several other species; Van der Horst Noble and Weber, '29; Fawcett and Hilfer, et al., '91, on the ultrastructure of the sperm '61; Picheral, '67; Nicander, '70; Martan and of Typhlonectes natans). Only a few of these Wortham, '72; Reed and Stanley, '72; Furi- reports include comparison of patterns of eri, '75; Wortham et al., '77, '82; Van der development and morphology across species, Horst, '79; Mainoya, '81; Mo, '85; Mizuhara et al., '86; Visser and Van der Horst, '87; Address reprint requests to Marvalrr H Wake, Departmerit or Yoshizaki, '87; Garrido et al., '89). Franzen Integrative Biology, University of Califnrnia. Berkeley, (:A 94720 o 1994 WILEY-LISS,INC. 262 M.H. WAKE and all are at the light microscope level (save were examined without a coverslip in bright- that of Van der Horst et al.). field illumination. I describe the comparative morphology of The maceration, several resuspensions, and mature sperm of a large number of genera especially the rinses in water cause loss of representing all five families of caecilians and numbers of sperm, but other nonsperm cells comment on the range of morphology, bioge- also are removed. Thus searching the field ography, and life history found in the order. I usually produces several sperm that are well examine both quantitative and qualitative stained and isolated from each other. Despite variables across all species and for a large years of preservation and the relative crudity sample within and among individuals of a of the technique, the sperm are quite intact. single population in order to assess patterns Even the acrosome, which I would have ex- of variation. The currently accepted hypoth- pected to be lysed and absent, was present in esis of the phylogenetic relationships of fami- virtually all specimens. However, the fragile lies of caecilians (Duellman and Trueb, '86; undulating membrane of the flagellum was Nussbaum and Wilkinson, '89; modified by present in only a few taxa and best repre- Hedges et al., '93 to include five families, sented in those most recently preserved (al- rather than six, the "Typhlonectidae" now a though it was also present in one long- subfamily of the Caeciliaidae) has been used preserved-70 yr-sample) . I suspect that as a basis for assessing phylogenetic patterns an undulating membrane may be present in presented by features of sperm morphology. more taxa than those that I observed (see MATERIALS AND METHODS Discussion), so I did not include its presence One to three testis lobes were removed or absence in my cluster analysis. The unim- from preserved specimens in various muse- paired morphology of long-preserved sperm ums (see Table 1and Appendix for list of taxa was evident when light microscopy photo- and their sources). Testis lobes vary in size, graphs of sperm of Typhlonectes natans were shape, and number across taxa (Wake, '68). compared with the ultrastructural detail pre- The lobes were macerated rapidly with dis- sented by Van der Horst et al. ('91).General secting needles in a watch glass to distribute structural fidelity was striking. Further, the sperm into physiological saline solution (0.7% measurements of sperm components ob- NaCl). A modification of the technique of tained are quite similar to those reported by Howell and Butts ('83) for silver-staining Seshachar ('39, '42a, '43, '45) of material sperm extracted from recently dispatched, prepared by sectioning and staining and by unpreserved fishes was used to obtain stained Van der Horst et al. ('911,using transmission sperm. The technique works well for pre- electron microscopy, for the same species. served amphibians, but length of time since The technique allows more details of acro- preservation does affect the quality of the some morphology to be revealed than those sperm produced (see below). reported by Seshachar ('40). Sperm suspensions were centrifuged 5 min Slides were photographed with a Nikon at 300 g; saline was aspirated and the sperm photomicrographic apparatus at 1,000 x with button was resuspended in 4 ml of 10%forma- oil immersion. Photographs of a micrometer lin. This procedure was repeated twice. Then, stage scale were taken periodically to be sure the supernatant was discarded and the sperm that photographs were taken at the same were resuspended in 1 ml of 10% formalin. magnification and that prints were all made Two to five drops of sperm suspension were to the same scale. Measurements were taken placed on a slide, and the suspension was by placing photographs on a digitizing tablet. distributed evenly. Slides were air-dried, The scale was calibrated, and (1) acrosome rinsed in running deionized HzO, and blotted length, (2) acrosome width, (3) head length, dry. Two drops of colloidal developer solution (4) head width anteriorly, (5) head width (2 gm USP gelatin dissolved in 100 ml deion- posteriorly, (6) body (midpiece) length, (7) ized HzO and 1ml formic acid) and 4 drops of body width, (8) flagellum length, (9) flagel- 50%AgNO3 (4 gm AgNO3 dissolved in 8 ml lum width anteriorly (at the proximal attach- deionized HzO) were pipetted to the surface ment to the body, and (10) flagellum width of the slide; the slide was covered with a 24 x posteriorly (at the distal tip) measured. Use 60 mm coverslip and dried on a 70" slide of the digitizer allowed the length of a long warmer for 3 min. The stain and coverslip and multiply curved flagellum to be mea- were then rinsed off in running deionized sured accurately. If it was apparent that a HzO, and the slide was blotted dry. Slides sperm head and body were not flat (i.e., curved COMPARATIVE MORPHOLOGY OF CAECILIAN SPERM 263 out of plane), so that shape was distorted in proportion to length). Long heads are >35 the two-dimensional photograph or the en- pm, moderately long 25-29 p.m, midlength tire flagellum was not in the photograph, the 15- 24 pm, short 8-14 pm. Narrow heads measurements were discarded. The paucity are 1 pm or less wide, medium width 2-3 pm, of sperm on the slides owing to the macera- broad > 4 pm. Midpieces considered long and tion and washing technique, coupled with thin are > 12 pm long and < 4 pm wide; long attempts for accuracy of measurements, re- but thick, > 12 km long but > 4 pm wide, sulted in very small sample sizes for most short and thick, 8 pm or < long and > 4 pm taxa. Meristic data are presented in Table 1. wide; short, slender, 8 or < pm long and < 4 Measurements were subjected to a princi- pm wide; moderate length 8-10 km. Flagella pal components analysis based on covari- are short if <lo0 pm, medium in length if ance. Several analyses were run: all samples 100-120 p.m, long if > 120 pm. for all species, including the large Dermophis mexicanus sample; one specimen per species Family Rhinatrematidae for all taxa; and the large D.