Methods

Study Area and Sampling Sites Utilized

         Prior to initiation of the USGS biological component of the planned IOS-NEGOM, the MMS had designated a target study area for the overall program of integrated investigations. This target IOS-NEGOM study area is defined as the large irregular polygon shown in Figure 1. Deep reef-fish community structure investigations undertaken by USGS from 1997-1999 (Weaver et al., 2002) were confined to the "Pinnacles" reef tract within this polygon, but the USGS area of interest was extended eastward beginning in 2000. Thus, samples for life history analyses were obtained from a larger geographic area than the original MMS IOS-NEGOM polygon. Small comparative collections from one western Atlantic site off North Carolina, and from the Florida Keys, supplemented material from the primary NEGOM sampling sites (Table 1).

Sample Collection

         Most material of the four study species was collected during USGS cruises to the outer-continental shelf reefs of the NEGOM from 1997 through 2002 (Figure 2, Table 1). Cruise and reef locations pertaining to geographic variations in P. martinicensis size-at-age are described in Tables 1 and 2, respectively.  Names of reefs used are not official geographic names, but place names of convenience designated by USGS researchers.  Fish were primarily collected by hook and line sampling while anchored over reefs. Small multiple unbaited hook rigs (e.g., Hayabusa Sabiki^®, and Mustad Piscatore^® rigs) were dropped to the bottom with weights (6-16 ounces depending on water current and depth), fished vertically just off the substrate, and jigged to imitate zooplankton. A variety of miniature hook sizes (nos. 3, 4, 6, 8, 10) were used; all hook sizes were effective in collecting all three serranid study species.  Only the smallest hook sizes were effective for C. enchrysurus.  The same multiple-hook rigs were also fished baited with small bits of squid; these were effective in capturing S. phoebe. A 4.9 m semi-balloon trawl with a 3.8 cm mesh body and a 0.6 cm mesh liner was also used to collect specimens inhabiting soft sediments and the reef rubble apron adjacent to the reef. The trawl was equipped with heavy rock-hopper wooden doors, a "Texas" roller rig consisting of a chain surrounded by 7.6 cm rubber disks hanging across the lower edge of the open trawl mouth, and a heavy coated chafing skirt.  This type of rock-hopper rigged trawl was used to reduce hang-ups and trawl damage while sampling over rubble substrates. A remote operated vehicle (ROV) (Phantom S2, Deep Ocean Engineering) equipped with a rotenone pump and suction tube sampler (38.1 mm inner diameter) (Weaver et al. 2002) was also used to collect smaller specimens.

         Supplementary study material of small P. martinicensis was obtained from the Florida Keys and the North Carolina OCS (Table 1). In the report, they were used solely for increasing sample size of small fish for the length-weight regression. Samples were worked up immediately for four cruises (TM 2001-01, TM 2001-02, TM 2002-01, and DMA-02) with both otoliths and gonads taken from most fish.  Otoliths were excised, cleaned, and stored dry. Gonads were removed, preserved in 10% formalin, and returned to the laboratory.  During the remaining cruises (TM 1997-01, TM 1998-02, TM 1999-01, TM 1999-02, TM 1999-03, TM 2000-01, TM 2000-02, CSA-03, and JSL-01) most of the specimens collected were frozen, and the remainder preserved in 10% formalin.  Sagittal otoliths were removed from the frozen specimens and gonads were excised from the formalin preserved specimens in the laboratory. Total body weight was taken from fresh or frozen specimens and weighed to the nearest 0.1 g.  Gonad weight was measured to the nearest 0.01 g from gonads fixed in formalin.  All lengths were measured to the nearest millimeter (mm).

Laboratory Methods

Age Structure

         In order to validate aging procedures, a subset (n=100) of P. martinicensis was aged using both whole and sectioned sagittal otoliths. Transverse sections (500 μm thick) taken from the left otolith with a low speed saw (Buehler Isomet^®) were placed on glass slides and mounted with Flotexx^® (Fisher Scientific). Whole right otoliths were immersed in water, distal side up to reduce glare.   Sectioned and whole otoliths were viewed through a dissecting microscope using reflected light.  Age estimates (number of concentric opaque rings counted on otoliths) generated from whole versus sectioned otoliths showed a 96% agreement in P. martinicensis.  There was no difference (paired t-test, =0.05) in ages between the two methods used.  The four fish receiving different age estimates were different only by a single annulus located on the margin.  Because no difference was detected, subsequent ages of all serranid species were determined solely from whole otoliths (methods described above). Whole C. enchrysurus otoliths were determined to be unsuitable for aging.  No ring structure was visible under reflected or transmitted light.  As a result, C. enchrysurus otoliths were sectioned and aged as described above. Samples of all species for which ages could not easily be estimated were omitted from analyses.

         Annual growth rings (annuli) are regions of denser aragonite deposited during periods of slow growth, January through April (Secor et al. 1991). As growth slows in winter months, otolith growth slows and the result is an opaque band observed under reflected light. The ring structure observed on P. martinicensis otoliths was confirmed to be annular (deposited annually) with a marginal increment analysis (Geffen 1982).  To validate these rings as annual structures, only P. martinicensis with a single, completely formed annulus were used for validation.  The distance from the otolith core to the outer edge of the rostrum was measured to the nearest 1.0 μm using a video camera + monitor system and image analysis software (Optimas 6.0^®) attached to a dissecting microscope. The distance from the otolith core to the outer edge of the annulus was subsequently measured, followed by the distance from the outer edge of the annulus to the outer edge of the otolith (marginal increment). Marginal increments were averaged monthly and plotted to determine the time of formation and the number of rings formed in a year.  To aid in determining the duration of annulus formation, the location of the outermost annulus was recorded (i.e., is the annulus completely laid down or still being formed?). If this last annulus was on the edge of the otolith (i.e., still being formed), it was not counted. Ring structure on the remaining study species was unable to be validated as annuli due to limited samples during the Fall and Winter months.  For these species, ring structure is assumed to correspond to age (years).

         Annual survival rate was estimated from abundance of individuals collected in each year-class (Robson and Chapman 1961). Annual survival was determined by this formula:

Annual Survival Rate =     T    
                                       n + T-1

Where:
 n = N₀ + N₁ + N₂ + N₃ + …

N₀ = number of individuals in youngest age-class fully sampled.

N₁ = number of individuals in next   oldest age-class, etc.

T =N₁ + N₂ + N₃ + …

         The modal age-class of each species and older individuals were included in the survival estimate to minimize size biases in sampling methods.

Reproductive Biology

         Gonads were preserved in 10% phosphate-buffered formalin for a minimum of four weeks. Gonads were rinsed out of formalin by soaking twice in tap water for one hour and a third time for at least 12 hours, before being transferred to 70% ethyl alcohol (ETOH) for storage and histology.  Preserved gonads were dehydrated in a series of increasing concentrations of ETOH, embedded in paraffin, sectioned along the transverse plane (4 μm thick), and stained with hematoxylin and eosin. Hermaphrodism in each species was defined following Sadovy and Shapiro (1987). Both P. martinicensis and H. vivanus are protogynous hermaphrodites, where individuals hatch as females and subsequently transition to males.  Individuals of these two species were assigned a developmental stage based upon the most advanced oocyte stage present for functional females (perinucleolar, cortical alveolar, vitellogenic, nucleus migration, and nucleus breakdown) or the amount and proliferation of tailed sperm in functional males.  Transitional specimens, individuals with gonads undergoing sexual transformation, were characterized by the presence of tailed sperm and non-regressed or regressing oocytes. Temporal variations in the mean size-at-age of P. martinicensis and of H. vivanus were made by comparing our samples to those of Coleman (1981) and Hastings (1981), respectively.  Serranus phoebe is a simultaneous hermaphrodite and was staged as both a male and female when both tissue types were present in histological samples. Macroscopically, testicular material in this species is creamy white and is located on the anterior periphery of the gonad, as well as in lobes protruding into the interior of the gonad. Ovarian material is yellowish in color and dominates the posterior two-thirds of the gonad.  Testicular tissue was occasionally omitted from histology samples of S. phoebe if the transverse section had been taken too close to the posterior end of the gonad. If testicular tissue was missing, S. phoebe was staged solely as a female. For all species, female sexual maturity was determined during the spawning season by the presence of vitellogenic oocytes. Vitellogen is a protein secreted by the liver. It accumulates in yolk globules in the cytoplasm of oocytes when the female is reproductively active (Wallace and Selman 1978). During resting periods, female maturity was determined by the presence of atretic oocytes, indicating prior spawning had occurred. In addition, the presence or absence of post-ovulatory follicles (POFs) was recorded.

         Reproductive periodicity was determined using the POF method (Hunter and Macewicz 1985).  Post-ovulatory follicles were assumed to become indistinguishable from other atretic bodies after approximately 24 hours, so their presence as collapsed structures with identifiable thecal and granulosa layers indicated that an individual female had spawned within a day of capture.

Fecundity Estimates

Gonads containing hydrated oocytes were used to assess the reproductive output for a single spawning event (batch fecundity). The left lobe of each P. martinicensis gonad was sectioned for histological examination while the right lobe was used for batch fecundity. Lobe fecundity was estimated by separating hydrated oocytes from the tunic using forceps and a low-pressure water stream in a Petri dish.  Once hydrated oocytes were isolated, they were stored in 4% phosphate buffered formalin.  Hydrated oocytes were placed in a Petri dish, covered with a 2:1 water:glycerin solution to prevent desiccation, and counted with under a dissecting microscope.  Lobe fecundity was used to extrapolate batch fecundity using the formula:

Batch Fecundity =

   Lobe Fecundity* Total Gonad Weight
                         Lobe Weight

Where:

Lobe Fecundity = average of six counts of number of hydrated oocytes present in a single gonad lobe.  Replicates varied less than 4 oocytes (<2%)/lobe.

Total Gonad Weight = total weight of the whole gonad (0.001 g).

Lobe Weight = total weight of right gonad lobe (0.001 g)

         The gonad structure of S. phoebe prevents traditional fecundity methodologies from being used. These gonads are undelimited (male and female tissue is not separated by connective tissue) and the two types of germinal tissue cannot easily be isolated.  No hydrated H. vivanus were collected during the study.