Loefer and Sedberry: Life history of Rhizophonodon terraenovae off the southeastern United States 



77 



The ovaries and uteri of females were examined macro- 

 scopically for indicators of maturity, such as yolking eggs, 

 embryos, or placental scars. Vitellogenic oocytes were eas- 

 ily identified by their bright yellow coloration in contrast 

 to the pale white coloration of nonvitellogenic oocytes. If 

 vitellogenic oocytes were present, the diameter of all vitel- 

 logenic oocytes in the ovary was measured (to the nearest 

 0.1 mm) with dial calipers. If maturing oocytes were not 

 present, the most differentiated nonvitellogenic oocytes 

 (which were noticeably larger that the rest of the oocytes 

 in the ovary) were measured. Any embryos were removed 

 from the uteri, counted, their sex determined, and mea- 

 sured (TL). Female maturity was determined by the 

 presence of embryos, umbilical scars in the uterus from 

 previous pregnancy, or the presence of large vitellogenic 

 oocytes (greater than 15 mm diameter) nearing ovulation 

 (Parsons, 1983b). 



A segment of the vertebral column extending from 

 the cervical region (dorsal to the branchial chamber) to 

 the origin of the first dorsal fin was removed from each 

 specimen and frozen. Vertebrae from the cervical portion 

 of the spinal column were used for aging because of the 

 shallow concavity of the intermedalia and the size simi- 

 larity between adjacent centra in this region. The shal- 

 low concavity of the vertebrae facilitated processing and 

 measurement during aging (Branstetter and McEachran, 

 1986). Age determination was attempted on 890 of the 

 1093 specimens collected during the study. Vertebrae 

 selected for aging were separated from the frozen seg- 

 ment, defrosted, and soaked in 5% sodium hypochlorite for 

 5-30 min (depending on size) and were removed from the 

 solution as soon as all excess connective tissue had been 

 dissolved. A longitudinal section approximately 500 pm 

 thick was cut from the center of each vertebrae with a Mark- 

 V watering saw and allowed to air-dry for at least 24 h. 

 Dried sections were then attached to glass slides with 

 Accu-mount 60 mounting medium and hand polished with 

 wet 600-grit sandpaper to a thickness of approximately 

 350 pm. Several staining or ring elucidation techniques 

 (e.g. Parsons, 1983a; Branstetter, 1986; Brown and Gru- 

 ber, 1988; Hoenig and Brown, 1988) failed to significantly 

 increase increment visibility; therefore all aging was per- 

 formed with unstained vertebral sections. 



Vertebral sections were read on a dissecting microscope 

 with transmitted light and a polarizing filter at 20x mag- 

 nification. Increment radii and marginal increments were 

 measured through the center of the corpus calcareum (Fig. 2) 

 with OPTIMAS image analysis software (Media Cyber- 

 netics, 1999). Precaudal length was regressed on centrum 

 radius (CR) for males and females to test for an isometric 

 relationship. 



The increments observed in vertebral sections were 

 narrow circuli similar to those described by Simpfen- 

 dorfer (1993), as opposed to the gi-owth bands described 

 by Branstetter (1987a). All increment counts were made 

 without knowledge of the size, sex, or collection date of 

 the specimen. The primary reader (senior author) counted 

 increments on all samples twice; each reading was sepa- 

 rated by at least two months. Increment counts that were 

 not in agreement were counted a third time. If the third 



Figure 2 



Diagrammatic representation of a vertebral sec- 

 tion; bm = birth mark, c = circuli, cc = corpus 

 calcareum, cr = line of centrum radii and annuli 

 measurements, f = focus, i = intermedalia. 



count did not agree with one of the first two counts, the 

 specimen was excluded from the analysis. The secondary 

 reader (coauthor) counted increments from all specimens 

 not eliminated by the primary reader's analysis. Between- 

 reader disagi'eements were re-examined by both observ- 

 ers simultaneously. All specimens for which a consensus 

 could not be reached were discarded. The index of average 

 percentage error (lAPE; Beamish and Fournier, 1981) was 

 used to estimate precision between the final readings of 

 the primary reader and the initial readings of the second- 

 ary reader 



The annual periodicity of increment formation was 

 verified through marginal increment analysis and focus- 

 to-increment frequency distributions. Absolute marginal 

 increment distances were converted to "relative" marginal 

 increments by dividing the distance between the last in- 

 crement and the edge of the centrum by the width of the 

 last fully formed growth band (Skomal, 1990; Natanson, et 

 al., 1995). This conversion compensated for differences in 

 growth rates between age classes. 



Back-calculated lengths at previous ages were esti- 

 mated from vertebral measurements by using a modified 

 Fraser-Lee equation proposed by Campana ( 1990): 



L„ = L,, + |(C„ - C,.) (L,. - L„)/{C,, - Cf,)], 



where L^ = length at age; 



L_ = length at capture; 



C^ = centrum radius from focus to increment a; and 



C = centrum radius at capture. 



