Size-Specific Vulnerability to Predation and Sensory System Development ofWhite Seabass, Atractoscion nobilis, Larvae
Daniel Margulies
ABSTRACT: The size-specific vulnerability of strength. Predation has long been recognized as white seabass, Atractoscron nobilis, larvae (2.5-15.0 an important potential regulatory mechanism in mm SL) to two types of fish predators-adult prerecruit life stages (Houde 1978, 1987; Hunter northern anchovy, Engraulis mordax, and juvenile 1981, 1984), but many of the specific factors that white seabass-was examined in laboratory preda influence predation rates on early life stages are tion trials. Concurrent analyses were made of the not well understood (Bailey and Houde 1989). developmental ontogeny of larval visual and mech anoreceptive systems. The proportion of larvae re One of the crucial aspects of predator/prey sponding to or escaping attacks by either predator dynamics involves the developmental or physio increased with larval size. There were no signifi logical basis for capture and escape responses. cant differences in proportions oflarvae responding Fish larvae possess several sensory systems pre to or escaping attacks of either predator until sumed to be important in predator detection; larvae were 6.0-7.5 mm SL (early postflexion these include visual, mechanoreceptive, and stage). At this size, larvae were better able to evade auditory systems. During early ontogeny, these the slower, more discontinuous attacks of juvenile sensory systems undergo rapid development, white seabass compared with the faster attacks of and the improvement in these systems is adult anchovy. Major developmental events occur thought to be crucial in controlling vulnerability during this larval stage, including rapid improve to predation (Hunter 1984; Blaxter 1986). How ment in visual acuity, visual accommodation to dis tant objects, growth and stratification of the optic ever, experimental investigations combining the tectum, and large increases in the numbers of free study of larval sensory system development and neuromasts on the head and trunk. It is likely that vulnerability to predation are limited. Larvae of at larval sizes >7 mm SL, the slower attacks of several species, including the northern anchovy, juvenile white seabass allow more time for larvae to Engaulis 'nwrdax, (O'Connell 1981; Webb 1981; process and integrate visual and mechanoreceptive Folkvord and Hunter 1986), Atlantic herring, sensory input from several modalities. White sea· Clupea harengus harengus, (Blaxter et al. 1983; bass larvae (and sciaenids in general) become more Blaxter and Batty 1985; Fuiman 1989), Cape demersal in distribution during the late larval anchovy, Engraulis capensis, (Brownell 1985), stages in nearshore southern California waters. bloater, Coregonus hoyi, (Rice et al. 1987), and This ontogenetic shift deeper into the water col· white perch, Morone americana, (Margulies in umn may be related to predator-avoidance capabil ities ofthe larvae, since most demersal planktivores press) have been studied to examine either sen exhibit some type of hovering, ambush, or discon sory system development or predator-avoidance tinuous mode of predation similar to that of juve behaviors. Similar studies have been conducted nile white seabass and attack at much slower on larvae of flatfish (Neave 1984, 1986) and sev speeds than pelagic, shoaling fish predators. As eral gadoids (Fridgeirsson 1978; Bailey 1984). they shift downward, older white seabass larvae These types of investigations, however, have not may maximize their predator-detection capabilities been combined during one study using a single when encountering demersal fish predators. cohort of experimental fishes. The purpose of this study was to examine, For most marine and estuarine fishes, preda experimentally, the size-specific vulnerability of tion-induced mortality during early life stages white seabass, Atractoscion rwbilis, larvae to may be crucial in determining year-class different types of fish predators, and to deter mine if there was a developmental or neuro Daniel Margulies, Southwest Fisheries Center La Jolla logical basis for any observed changes in larval Laboratory, National Marine Fisheries Service, NOAA, vulnerability or behaviors. The developmental P.O. Box 271, La Jolla, CA 92038; present address: Inter American Tropical Tuna Commission, c/o Scripps Institution studies of larval white seabass centered on the of Oceanography, La Jolla, CA 92093. ontogeny of two sensory systems, vision and
Manuscript Accepted April 1989. 537 Fishery Bulletin, U.S. 87: 537-552. FISHERY BULLETIN: VOL. 87, NO.3. 1989 mechanoreception, which are presumed to be southern California and are potential natural important components of predator detection and predators of white seabass larvae. These two larval escape responses (Blaxter 1986). The predator groups represent different types of ontogeny of avoidance/escape responses and raptorial predatory behavior (Table 1). Adult other larval behaviors of white seabass were northern anchovy are pelagic, cruising predators then analyzed in relation to the structural and that attack prey at relatively high speeds and functional development of their sensory often from oblique angles. Juvenile white sea systems. bass attack at much slower speeds and have a Very little is known of the early life history of somewhat discontinuous mode of attack that in the white seabass, even though it is the largest volves an approach, glide, and then engulfing of sciaenid occurring in coastal waters of California prey. Northern anchovy also initiate attacks and Baja California and has historically been a from much greater distances. favored sport and commercial species (Feder et Predators were held in laboratory holding al. 1974; Vojkovich and Reed 1983). White sea tanks and fed adult Anemia, minced euphausiids bass spawn in spring and early summer in near and occasionally white seabass larvae. To mini shore coastal waters of southern California, and mize predator learning behavior, after comple produce pelagic eggs. Abundance and distribu tion of predation trials on a given date, predators tion of young-of-the-year in southern California were transferred to separate holding tanks; this waters have been reported by Allen and Frank ensured that no predator was used more than lin (988). Larval morphological development twice during the 6 wk experimental period. has been described by Moser et al. (1983), but larval ecology, predator-avoidance behaviors, Experimental Methods and sensory system development have not been described. The predation experiments were conducted following the methods of Folkvord and Hunter (1986), with several modifications. Predation METHODS trials were carried out in rectangular fiberglass 3 Larval Rearing tanks, 1.4 m in volume, with a clear glass win- White seabass larvae used in the experiments 16 were hatched from eggs obtained from brood stock adults maintained at the Hubbs Marine A, 14 ,, Juvenile Research Center at Sea World in San Diego. , ,I Eggs were transported to the aquarium at the K Southwest Fisheries Center La Jolla Laboratory i 12 I I and stocked in 76 L aquaria at stocking densities ! I :c I of 6 L- 1. During the culture experiments l- Cl 10 i water temperature ranged from 17° to 19°C, and z I W I a constant photoperiod of 13 h light: 11 h dark ..J I Q ,f was maintained. Ill:c 8 Q ,f First-feeding white seabass larvae were fed z C rotifers, Brachionus pLicatilis, cultured on the I- green alga Tetraselmis. Older larvae were se en 6 ,,Y", Pastflexian quentially fed diets ofAnemia nauplii, a mixture ..Jc ,f, > ,." Flexion of Anemia nauplii and wild copepods, and adult Ill:c 4 Anemia. The developmental size range of white ..J ,t-" seabass tested ranged from hatching (2.6 mm , SL, 0 day after hatch) to juvenile metamorphosis 2 Hatch (15.0 mm SL, 42 days after hatch) (Fig. 1).
o 10 20 30 40 50 Experimental Predators LARVAL AGE (days after hatch) Adult northern anchovy and juvenile white
seabass were used as predators. Both of these FIGURE l.-Sizes, ages, and developmental stages of predator groups occur in nearshore waters of larval white seabass prey used in predation trials. 538 MARGULIES: VULNERABILITY AND SENSORY DEVELOPMENT OF WHITE SEABASS
TABLE 1.-Modes of predation exhibited by adult Engraulis mordax and juvenile Atractoscion nobi/is.
Attack Size (mm) distance (dm) Attack speed Predator Range Mean Range Mean range (m/s) Mode of attack Engrau/is mordax 85-107 96.4 20-70 36.6 0.50-1.50 Fast speed; approach and (Adults) attack continuous; often from oblique angle Atractoscion nobi/is 62--86 72.2 5-30 14.7 0.05-0.20 Slow to moderate speed; at (Juveniles) tacks often discontinuous, with approach, glide, then ingestion; angle of attack horizontal or oblique dow on one side for observation. Flourescent ture tanks and fixed in 5% formalin for calcula lamps produced 2,000--3,000 mc at the surface of tion of mean larval sizes; additional subsets of each tank. A black plastic tent enclosed the win larvae were removed for examination of sensory dow, providing a darkened compartment for system ontogeny (discussed below). The total observations. Metric rules placed on the tank number of predator-prey interactions observed window allowed estimation of predator attack for each larval size class and predator type is distances. The tanks were supplied with flow presented in Table 2. through ambient seawater 07°-19°C), except during an experiment when the water was static. Since adult northern anchovy often occur TABLE 2.-The total number of predator-prey interactions ob in large schools and juvenile white seabass are served for each larval size and age class and predator type. solitary or found in loose aggregations, it was X larval size Larval age Observations per predator type decided that northern anchovy predators in (mm) (d) Northem anchovy White seabass groups of five individuals and white seabass in 3.1 4 36 39 groups of three should be used. This simulated 4.5 12 35 32 the predators' natural condition yet prevented 6.1 17 33 38 predator "swamping" of larval prey. 7.4 23 30 37 Feeding motivation of predators was stan 9.1 27 40 38 dardized by presenting constant numbers of 12.5 34 38 40 adult Artemia to predators prior to predation 14.8 42 42 50 trials with larvae. (Arlern-ia are a good standard prey because they do not avoid predatory at tacks by fishes.) Preliminary experiments had Classification of behaviors followed Folkvord indicated that feeding behavior ofadult northern and Hunter (1986). Four measures of predator anchovy was more variable than that of juvenile prey interactions were calculated: mean and white seabass; thus, five groups of five Artem-ia maximum attack distances; percentage of larval each were presented to each anchovy predator avoidance responses; percentage of larval group and three groups of five Artemia each escapes; and predation rate (percentage of were presented to white seabass. After the larvae captured during each 10 min trial). Ap Arte?nia additions, white seabass larvae were proximate predator attack speeds also were introduced into the predation tanks in groups of estimated. A predator attack was a directed five (each larval group added = trial). Trial dura movement toward a prey with the mouth open. tion was 10 minutes or until all prey were eaten. Predator attack distance was the distance in Larvae and Arte'mia were added to the tanks in decimeters (dm) from a prey to the point of clear beakers that were gently submerged at the origin of attack. An avoidance response was a water surface. After the initial Arte'm;ia trials, change in speed or direction of a larva occurring each predator group was tested with 6--8 larval before a predator could complete an attack. An trials, followed by a final Artemia trial to test for escape occUITed when a predator failed to cap predator satiation. ture a larva during a single attack. Repeated On the same day as predation experiments, attacks were scored as separate events. subsets of 10--12 larvae were removed from cul- Two potential biases to the predator-prey 539 FISHERY BULLETIN: VOL. 87. NO.3. 1989 interactions were addressed. Variation in preda and then removed for examination of number tor performance was examined by calculating and location of free neuromasts under light the percentage of predator en'or occurring in microscopy. five groups of both types of predator feeding on Histological sections also were examined to adult Artem-ia. Since Artem;ia show no avoid determine the timing of swimbladder inflation in ance responses, a predator error was recorded larvae. when a predator simply missed the prey. Per centage ofpredator en'or was calculated for each Data Analysis predator group and among groups. In addition, potential stress or mortality of larvae due to For each predation trial, predation rate (per handling was examined. For each larval size cent killed in 10 minutes) was calculated from the class, four replicates (trials) of five larvae each equation: A = 1"n + 1'1 - mn, where A = total were transferred into 76 L tanks containing no mortality rate, rn = control mortality rate, n = predators. Mortality observed in control tanks experimental predation rate, and mn = the in was used to adjust for handling-induced mortal teraction effect which estimates the proportion ity or increased vulnerability of larvae in preda of larvae consumed but which would have died tion trials. anyway from handling stress or other causes (Ricker 1975). Percentage of larvae responding Analysis ofLarval Sensory System and escaping were considered total survival Development rates (8) and were calculated by first estimating total conditional mortality (A) (proportion "not Subsets of larvae from each size class were responding" or "not escaping") and then calculat removed from culture tanks for analyses of ing the difference as 8 = 1 - A (Ricker 1975). larval visual and mechanoreceptive systems. Lens diameter, visual acuity, and thickness of Groups of 6--8 larvae from each size class were the optic tectum were calculated for each larval fixed in 5% phosphate-buffered formalin for his size class and developmental stage. Develop tological analysis of organogenesis. Larvae were mental ontogeny of the retina, optic tectum, and dehydrated, cleared, embedded in paraplast, swimbladder were described. The mean number and serially sectioned at 5 IJ.m transversely and of neuromast organs also was calculated, and sagittally. Sections were stained with Harris' composite maps were developed of the patterns haemotoxylin, counterstained with eosin, and of neuromast organ formation. viewed under light microscopy at 80--1250 mag Statistical analyses of data were performed nification. Ontogenetic development of the lens, using SAS (SAS 1982) statistical programs. retina, and optic tectum were examined and swimbladder inflation was noted. Size-specific visual acuity was calculated from retinal sections RESULTS using the formula: sin a = elf, where a is the minimum separable angle, e is the distance be Predator/Prey Interactions tween the centers of adjacent cones, andfis the Potential Biases focal length ofthe lens (Neave 1984). Cones were measured as numbers per 100 IJ.m length of Preliminary determinations of predator error retina (d), thus the reciprocal 10 d gives cone by both predator types indicated that predator separation in mm. The expression wao multiplied performance would not bias results. Mean preda by 1.11 to adjust for approximate 10% shrinkage tor error per trial for northern anchovy adults during processing, and the focal length of the feeding on Arte1"nia was 3.3% (range 2.1--5.1%) lens was calculated by multiplying its radius (r) while error rate for juvenile white seabass was by 2.55 (Matthiessen's ratio; Matthiessen 1880), 2.7% (range 0.8--6.1%). There were no signifi giving: sin a = 0.0435/dr. cant differences in mean error rate among preda Separate groups of 6--8 larvae from each size tor groups, within trials, or between species class were examined for development of free (ANOVA, P > 0.10). neuromast organs. These larvae were anaesthe Larval mortality in control tanks (no preda tized with MS-222 and immersed in a bath of the tors) was low, ranging from 0.0 to 12.5% mean vital stain Janus Green (0.05% Janus Green values for any larval size group. Experimental made up with 50% seawater) (Blaxter et al. predation and escape rates were adjusted by the 1983). Larvae were immersed for 20--30 minutes control rates. 540 MARGULIES: VULNERABILITY AND SENSORY DEVELOPMENT OF WHITE SEABASS (Fig. 3B). Responses and escapes from juvenile Larval Responses white seabass improved significantly in larger The probability of a white seabass larva re larvae, particularly between the larval sizes of sponding to or escaping a predatory attack gen 7.5 versus 9.0 mm SL (t-test, P < 0.01). Re erally increased with larval size, although major sponse and escape success nearly doubled during differences in these responses were apparent de this developmental stage. Mean predation rate pending upon type of fish predator. The mean by juvenile white seabass predators decreased percentage of larvae escaping attacks by north from 80% on yolk-sac larvae to 65% on 6 mm ern anchovy ranged from 3% in yolk-sac larvae (3 larvae, increased to 87% on 7.5 mm larvae, and mm SL) to 43% in metamorphosing fishes (15 then steadily decreased to 30% on early juveniles mm SL) (Fig. 2A), while the percentage re (Fig. 3A). These predation functions also were sponding to anchovy attacks ranged from 14% fit to exponential regressions (Fig. 3). (yolk-sac stage) to 56% (at metamorphosis) (Fig. The success of avoidance movements by re 2B). Mean predation rate (percentage of larvae sponding larvae (numbers escaping/numbers re eaten in 10 minutes) by anchovy predators de sponding) generally increased with larval size creased from 95% on yolk-sac larvae to 56% on (Table 3). Approximately 21% of responding early juveniles (Fig. 2A). The predator/prey larvae in the 3.1 mm size category successfully interactions were best described by exponential escaped northern anchovy attacks; this percent regressions (Fig. 2). age increased to 50% in 4.5 mm larvae and 75% White seabass larvae were better able to at metamorphosis. Avoidance success from juve respond to or escape attacks of juvenile white nile white seabass attacks ranged from 39% for seabass than those of northern anchovy. Mean 3.1 mm larvae to 75% for 7.0 mm individuals and percentage of larvae escaping white seabass at improved to nearly 90% in early juveniles. tacks ranged from 8% for yolk-sac larvae to 74% Statistical comparison of the responses and for metamorphosing fishes (Fig. 3A), while the escapes from northern anchovy versus white percentage responding to attacks increased from seabass predators indicated that a significantly 18% (yolk-sac stage) to 84% (at metamorphosis) higher percentage of larvae >6.0 mm SL
.... 100 Predator: Engrau/ls mordax .....'f1I. A ~ CJ ~ 80 I-+-+',J" e 1 ,1 + 60 CJz 1 ""1 i[ e ~ 40 /~ w w ~ 20 -t-r--f FIGURE 2.-Vulnerability of white seabass larvae to Ill: adult northern anchovy predators as a function of lar- :5 0 val length. A. Solid circles are percentage of larvae B escaping an attack, error bars are 2 x SE, regression CJ equation is Arcsine Y = 5.078 eO.I44SL (on = 44, or = 80 zis .... 0.73), where Y = proportion oflarvae escaping and SL zf. = larval standard length (mm); open circles are per- O~ 60 centage oflarvae eaten in 10 min trials, error bars are A.CJ 2 x SE, regression equation is Arcsine Y = 81.326 036SL (n ;e e-O. = 44, or = 0.96), where Y = propor- ~e 40 tion of larvae eaten in 10 minutes and SL = larval ~~ standard length (mm). B. Percentage of lan'ae Te- e sponding to an attack, error bars are 2 x SE, regres- ... 20 sion equation is Arcsine Y = 15.293 eO.OSlSL In = 44, r = 0.89), where Y = proportion of larvae responding and SL = larval standard length (mm). 0 2 4 6 8 10 12 14
LARVAL STANDARD LENGTH (mm)
541 FISHERY BULLETIN: VOL. 87. NO.3. 1989 .....j,i ~ 100 () A ce t: 80 oC CJ z 60 a: ce () tJ) 40 IU IU ce 20 >a:: FIGURE 3.-Vulnerability of white seabass larvae to juvenile white seabass predators as a function of lar :5 val length. A. Solid circles are percentage of larvae B escaping an attack, error bars are 2 x SE. regression CJ Z 80 equation is Arcsine r = 10.976 eO.120SL (n = 44, r = Eii 0.89). where Y = proportion oflarvae escaping and SL z .... = larval standard length (mm); open circles are per ~ ~ 60 centage of larvae eaten in 10 min trials, error bars are mce 2 x SE. regression equation is Arcsine Y = 81.210 lUI- e-O.04SSL (n = 44,r = 0.92), where r = propor :!C 40 tion of larvae eaten in 10 minutes and SL = larval ceo standard length (mm). B. Percentage of larvae re ~I- ce 20 sponding to an attack, error bars are 2 /. SE. regres ... sion equation is Arcsine Y = 16.822 eO.096SL (n = 44,/;!. = 0.94), where Y = proportion of larvae responding and SL = larval standard length (mm). o 2 4 6 8 10 12 14 LARVAL STANDARD LENGTH (mm)
Larval Visual Ontogeny TABLE 3.-Avoidance success of responding larvae (numbers escaping/numbers responding). The Retina and Visual Acuity Predator type Histological examination of the white seabass Northern anchovy White seabass visual system revealed numerous developmental changes from hatch to juvenile metamorphosis. Larval SL i escape iescape (mm) success SE success SE In yolk-sac larvae, the lens was present, but the l'etina was unpigmented and undifferentiated. 3.2 21.4 4.9 38.8 7.1 4.3 47.0 11.5 57.9 9.1 At time of yolk absorption (ca. 3.2 mm SL), the 6.8 63.0 13.2 70.0 7.0 retina became differentiated into distinct layers, 7.4 51.7 6.7 80.0 7.5 the photoreceptive layer contained visual cells 9.2 58.8 7.3 81.1 6.6 and the epithelial (basal) layer became pig 12.4 64.3 5.3 90.0 5.1 mented. Presumably, at this stage the eye was 14.8 75.0 6.0 88.0 4.5 functional. In young larvae (3.~7.0 mm SL), the retina appeared to be composed of only cone cells and escaped white seabass attacks, while a signifi no retinomotor FIGURE 4.-Sagittal section of the eye of a 4.8 mm SL white FIGURE 5.-Cross-section of the retina of an 8.0 mm SL seabass larva showing the position of the lens retractor white seabass larva showing several dark mitotic bodies (Ml muscle (LRl. (Al indicates the anterior direction. x 100. in the outer nuclear layer. x250. 543 FISHERY BULLETIN: VOL. 87, NO.3, 1989 u 100 r- ~ III A ...o 80 c ! 60 > l- S (,) 40 ce .J ce 20 :::) + en > 0 B .---. LENS DIAMETER - 700 ...... CONE DENSITY + - 60 0 E - ::I. 0 500 A 50 0 E \ 0,,% ,.- ... :::L - ...... -IX ··l\. 0 0 /t'- IS 1&1 400 f- .... , 40 c I ·····f.... 0 l' 1&1 -> ::E .... ,/ I- ce 300 ...;t...... 30 CiS Q • __ ' •.••• ,.,••0 Z " 1&1 en ~" ...... 00_..+ Q z 200 20 1&1 + 1&1 .J z , ,l 0 100 .,.~ 10 (,) 0 0 2 4 6 8 10 12 14 30 LARVAL STANDARD LENGTH (mm) FIGURE 6.-Visual development of larval white seabass as a function oflarval length. A. Visual acuity, error bars are 2 J< SE, regression equation is Y = 138 - 18.2 SL + 0.726 SL2 tn = 66, ,,2 = 0.96>, where Y = visual aruity (min of arc) and SL = larval standard length (mm). L.R. is the first appearance of the lens retractor muscle; Nuclei is the first appearance of mitotic bodies in the outer nuclear layer; and R.M. is retino motor response. B. Squares are changes in lens diameter, error bars are 2 )( SE, regression equation is Y = -58.7 + 37.8 SL (n = 78,r = 0.97), where Y = lens diameter (IJom) and SL = larval standard length (mm); circles lU'e changes in cone cell density, en'or bars are 2 x SE, regression equation is r = 52.6 - 2.3 SL (n = 66. ,,2 = 0.89), where Y = cone cell density (no./100 IJom) and SL = larval standlU'd length (mm). Open circles are measurements taken in the area. temporalis. Values for a 30 mm SL juvenile are given for comparison to larvae. FIGURE 7.-Retinomotor response of a 12.5 mm SL. light-adapted white seabass larva. showing migration of the epithelial masking pigment. ONL is the outer nuclear layer. x400. 544 MARGULIES: VULNERABILITY AND SENSORY DEVELOPMENT OF WHITE SEABASS III: 100 ILl A >c -I 80 III: ILl T.L. + z_ -t--"",-t zo 60 + ,t----t--- -0 11I:1(',.. ~·-·~- 1:- 40 r...... '" c , -I III: + ILl 20 ...::::l 0 0 300 B E e--... TOTAL ..:! 0---0 OUTER LAYER rn 250 t rn zILl -+----f----1 FIGURE 8.-Cross-section through the optic lIIIi: 200 ..--- tectum of a 9.2 mm SL white seabass larva. The U:c .r'" tectum is bilayered at this stage, with an inner ... 150 --~ stratum perit'entriculare (SPV) and an outer :Ii I, ~ _- ::::l stratum zonale (SZ). The SZ is thickening and ,..."'. fr-- -~---- - ...u 100 exhibits increasing numbers of neurons migrating ILl from the inner layer. The torus longitud'inalis (T) ... 1'/~.¢--4t/ is quite prominent by this stage. x !OO. 50 f '--''---'--J..---L.....L...... L...... L...... J...-.J..-.L...... J'---'--J..---L...... L-..r~ o 2 4 6 8 10 12 14 30 LARVAL STANDARD LENGTH (mm) length (Fig. 9B). Growth and differentiation of FIGURE 9.-Development of the optic tectum of larval white the SZ appeared to involve cell migrations from seabass as a function of larval length. A. Change in the ratio the inner SPY with a progressive increase in the ofthe outer layer/inner layer, T.L. is the first appearance of the toms longitudinalis. B. Solid circles represent total SZ:SPV ratio (Fig. 9A). This ratio increased thickness of tectum. error bars are 2 x SE, regression equa 2 rapidly at first feeding (-3-4 mm SL), stayed tion is :r = -48.6 + 38.8 SL - 1.35 SL (11 = 64,r = 0.97), constant at lengths of 4.0-7.5 mm SL, and then where Y = total width of tectum (ILm) and SL = larval increased again from 7.5 mm to early juvenile standard length (mm); open circles represent thickness of stages. outer layer, error bars are 2 x SE, regression equation is r = -33.8 + 20.3 SL - 0.56 SL2 (n = 60,,2 = 0.97), where Y Another important event in the ontogeny of = width of outer layer (ILm) and SL = larval standard length the optic tectum involved the development of the (mm). Values for a 30 nml SL juvenile are given for compari tarus longitu.dinaJis (TL). This structure first son to larvae. appeared histologically at a larval length of 6.8--7.0 mm SL (Fig. 9A). Seen in transverse section, the TL developed as a teardrop-shaped structure with a wide area of contact dorsad with the optic tectum and ventrad with the epithala mus (Fig. 8). The TL continued to grow with ontogeny, and in older larvae and early juveniles the TL exhibited an increasing number of neuronal projections to the cerebellum. 545 FISHERY BULLETIN: VOL. 87. NO.3. 1989 Ontogeny ofMechanoreceptors only other stained structures were the nasal pits. Neuromast Development The total number and patterned formations of Staining with Janus Green provided somewhat free neuromasts increased with ontogenetic de variable results, in that not all white seabass velopment (Figs. 10, 11). At first feeding, larvae larvae took up the stain equally well. Some had 6-8 neuromasts on the head and 6 or 7 on the neuromast organs stained better than others, trunk. During development, new head neuro but the use of 6-8 fish per larval size class masts appeared first in the supraorbital and in seemed to provide good composite patterns of fraorbital areas, while on the trunk they re neuromast formation. cruited midlaterally (Fig. 11). The period of most At hatching, larvae had 5 or 6 pairs of neuro rapid addition of free neuromasts was in the masts on the head and 5 or 6 pairs extending larval size range of4-9 mm SL. During the early along the trunk (Fig. lOA, B). All of these postflexion stage (-7-10 mm SL), free neuro organs appeared to be free neuromasts (not en masts began to form in distinct supraorbital. in closed by canals). Each free neuromast consisted fraorbital, and preopercular (hyomandibular) of a basal, naked neuromast organ attached rows on the head and in a single midlateral row apically to a cylindrical, gelatinous cupula. The on the trunk (Figs. 10, 11). By 12.5 mm SL, 60 / .. A ./ rn A/TOTAL ... 50 ...- rn ,/ c ,/ :E ...... 0 40 , a:: =w /l' ~~~~+ z 30 ~ II. .+ +...... -t... 0 ,~{ a:: .....·t······· , ,/'r;;;;K w 20 ID :E +t'-f ...... t",A---9"'/ = Z 10 .,.~"9'.J ,r + + + + 1.0... P.O. Partial L.L. H••d Partial Partial L.L. Canal. Dlallnct 0 2 4 6 8 10 12 14 16 60 rn 50 ...rn c :E 0 a:: 40 =w z 30 II. 0 a:: w 20 ID :E =Z 10 0 5 10 15 20 25 30 35 40 LARVAL AGE (days) FIGURE lO.-Numbers of free neuromasts developing on larval white seabass as a function of larval length (AI and age (8). Error bars are 2 x BE. 1.0.'s are infraorbital and supraorbital rows on the head; P.O. is the preopercular or hyomandibular row on the head; and L. L. is the lateral line. Values are based on staining and should be considered approximate only. 546 MARGULIES: VULNERABILITY AND SENSORY DEVELOPMENT OF WHITE SEABASS FIGURE 1I.-Composite maps showing the location offree neuromast organs during development of white seabass larvae. Larval outlines were modified from those of Moser et al. (1983). some of the recruiting neuromasts were being DISCUSSION enclosed by canals on the head and trunk; some Neurosensory Basis For Avoidance existing free neuromasts were being incor Responses porated in canals as well. At juvenile meta morphosis, the three major branches of the head The overall probability of white seabass larvae canals, as well as the lateral line on the trunk, escaping predatory attacks seems highly depen were fairly distinct (Figs. lOA, 11), forming the dent on the size of the larvae, the type and qual precursor to the juvenile and adult lateral line ity of sensory input being integrated by the system. larvae, and the type of predator encountered (Fig. 12). Yolk-sac larvae have nonfunctional eyes, a small number of free neuromast organs Swimbladder Development on the head and trunk, and a noninflated swim In yolk-sac larvae, the swimbladder appeared bladder. At fIrst feeding (-3.2 mm SL), the eye as a collapsed sac dorsad to the yolk sac, while in becomes functional, but visual acuity is poor; the fIrst-feeding larvae it was still partially flattened pure-cone retina limits peripheral vision and mo and situated dorsad to the anterior digestive tion detection, and no accommodation is possible tract. Timing of complete swimbladder inflation (since there is no lens retractor muscle). There is was variable; most larvae exhibited full inflation strong correlative evidence that increases in at lengths of4.f>-5.5 mm SL. In larvae >4 mm, a numbers of free neuromasts and improvements pneumatic duct was present in the anterodorsal in the visual system are responsible for the area of the swimbladder. It was not clear improved avoidance responses observed in pre whether swimbladder inflation occurred by way dation trials. In larvae <4.5 mm SL, predator of air-gulping at the surface or by gas secretion detection is most likely a function of mechano internally. reception by free neuromasts, since visual 547 FISHERY BULLETIN: VOL. 87, NO.3, 1989 Heleh Distribution More Demersel Juvenile Stage General Behavior • t Cennlballsm- - • I Late Y-sJFlrst-F..dlng Notochord Fllxlon F'lxlon COmpl... Swlmbladder (SB) + and Osteology t FullI • I SBPr...nl, olSBIn"" ... I NoIlnl\olod Po"1811.0•• S.O. Po.llol H.M. Row PIlrIlIILL. H••dlndLL I 5·7 Pair On He.... Trunk Rowo On Hood on Hood Soma Clinl' Orlllni Cana•• DI.llncl Free Neuromasts • + + • f ILen.PrI..nt DoubllCon••, I Eye FUnc\lonol ONL 1Iu11~TIo••d AculIY. Vision 1--IIolIII\.IpidIm...... nlln Aculty-l Acuhy'" T.L of Tectum MKolicBod'" RellnomolOr R..pon..(Rodl) IMllrlr: Celli Appelr. InONL Onty 100 ...1 _ % Responding .. " .. __e 75 -l Repld Improvement To White I .... _"" _1 _ ---_... ---- Seabass 50 -I Predators _..-----r---- 25 -I .. ----..----- o -+1------75 % Responding -I --_a Slow, Uniform Improvement • To Anchovy 50i Predators .... -- -_ .... ------25 1 .... --_ ..... --_ ...... -.. -- o '1------I o 2 3 54 6 7 8 9 10 11 12 13 14 15 Standard Length (mm) II II II 01 45 10 15 20 25 30 35 40 42 Days At 17-19'C FIGURE 12.-Developmental events and sensory system development during the early life history ofthe white seabass. Formatis based on Hunter and Coyne (1982). Developmental data are from my study, except some behavioral data from Orhun (unpubl. data) and osteological data from Moser et al, (1983). Y-S = yolk-sac stage, 1.0. = infraorbital, S.D. = supraorbital, H.M. = hyomandibular, L.L. = lateral line, T.L. = tOTltslongitudinaUs, ONL = outer nuclear layer of retina. acuity is poor and no acoustic inputs are likely visual improvement and integration. Visual through the swimbladder. During notochord acuity calculated for most adult marine fishes is flexion (-5--7 mm SL), visual accommodation is 2--10 minutes of arc (Tamura 1957); values for developed with the lens retractor muscle, the adult white seabass are unknown but probably swimbladder is inflated (potentially providing fall in this range. Thus, approximately 75--80% of acoustic stimuli), and there begins a major re the improvement in acuity seen from hatch to cruitment of free neuromasts on the head and adult stage in white seabass has occurred by the trunk. Up to this point in development, it would late larval stage. However, for vision to playa appear that vision plays a limited role in preda major role in predator detection, improvement tor detection. in acuity must be accompanied by development During the early postflexion stage (-7 mm), of rod cells, by accommodation to distant objects the visual system begins to undergo numerous (lens retractor muscle), and by growth and de changes (Fig. 12). Acuity continues to improve, velopment of the optic tectum. The development early rod precursors develop in the retina, the of rod vision helps to improve peripheral vision optic tectum increases markedly in size and (O'Connell 1981) and motion detection (Blaxter stratification, and the torus longitu,din..alis be 1986), both crucial aspects of predator detection. gins to develop in the midbrain. At 10.5--12.5 mm Rods also aid in improved visual performance in SL, double cones and early rod cells are present dimmer light (O'Connell 1981), which would in the retina. These changes are essential to improve foraging skills and predator-detection 548 MARGULIES: VULNERABILITY AND SENSORY DEVELOPMENT OF WHITE SEABASS as white seabass become more demersal. Growth uli detected through the gas-filled otic bullae and stratification of the optic tectum allow for seem important in the development of startle more complex interconnections with other brain responses of Atlantic herring larvae (Fuiman centers in teleosts and are essential for the de 1989). The inflation of the otic bullae with gas velopment of progressively more sophisticated and the occurrence of a well-developed acous behavioral responses (Munro 1984). The taru,s tico-Iateralis system, however, seems to be more longitu.dinalis functions in midbrain integration characteristic of clupeoid larvae (Fuiman 1989). of proprioceptive information (Groman 1982) and Although acoustic stimuli or Rohon-Beard is involved in the control of visual motor patterns (mechanoreceptive) input could also be related to (Munro 1984). improved evasion I'esponses of white seabass The free neuromasts in fish larvae probably larvae, the observed improvements in the function in detection of differences in velocity neuromast and visual systems seem to be between the fish and surrounding water. The directly related to the improved avoidance incorporation of neuromasts into canals (as oc capabilities. curs in older white seabass larvae and early juveniles) probably aids in schooling or detection of accelerations in water movements caused by Larval Vulnerability to Attacks other animals, such as predators (Blaxter et al. Larval size and developmental stage are the 1983). The increases in numbers and in pat most important factors related to larval vulner terned formations of neuromasts with white sea ability in laboratory trials. The type of predator bass larval size probably improve detection of encountered also influences predation rates. predator movements and aid in swimming move However, although white seabass larvae were ments and proprioception. better able to respond to juvenile white seabass During the early postflexion stage (7.5--10.0 attacks than those of northern anchovy, this did mm SL), these improvements in mechanorecep not result in significantly reduced predation tive and visual capabilities appear to be directly rates (in comparisons between predator types) related to improved detection and escape re until larvae were >8.5 mm in length. This sug sponses. However, depending upon the type of gests that other factors related to predator de predator encountered. a significant difference tection of prey, such as prey morphology, water exists in the degree of improvement in avoidance clarity, and alternative prey abundance, may be capabilities (Fig. 12). Early postflexion larvae as important as predator type in controlling vul exhibit a modest improvement in evading north nerability ofsmall white seabass larvae. Physical ern anchovy attacks, but display a dramatic im background and morphological conspicuousness provement in detecting and escaping the slower, of prey can be important factors controlling pre more discontinuous attacks ofjuvenile white sea dation rates of planktivorous fishes (Vinyard and bass. Since some startle responses are elicited O'Brien 1976), while relative abundance of al even in yolk-sac larvae, it appears that neural ternative prey has been shown to have signifi motor pathways such as Mauthner-type neurons cant effects on fish consumption rates on small (Eaton and DiDomenico 1986) are present and white perch larvae (Margulies in press). functioning during all developmental stages. One disadvantage of laboratory studies is that Just prior to and during the early postflexion realistic encounter rates between larval prey stage, larvae undergo notable additions ofneuro and fish predators are difficult to simulate. Al mast organs and major improvements in the vis though the main purpose of this study was to ual system. Since the outcome ofa predator-prey delineate the developmental basis for larval interaction is heavily dependent upon reaction avoidance behaviors, it is important to recognize velocity and timing (Webb 1976), the slower, the limitation of predicting total vulnerability of close-range attacks of juvenile white seabass larvae based on laboratory trials only. Total vul probably allow more time for detection of sen nerability to predation is a function of the prob sory stimuli from several modalities as well as ability ofencounters between predator and prey, sensory-motor integration needed for response the probability of capture of prey and the prob and escape movements. ability of attacks by a predator (O'Brien 1979). Improvements in visual and mechanoreceptive My data provide reliable estimates of probability systems have been implicated in the evasion be of prey capture and, to a lesser degree, probabil haviors of northern anchovy larvae (Webb 1981; ity ofattacks by the experimental fish predators. Folkvord and Hunter 1986), while acoustic stim- Encounter rates. however, are affected by a 549 FISHERY BULLETIN: VOL. 87. NO.3. 1989 number of larval growth-related parameters, in sciaenid larvae show a marked vertical size dis cluding increased larval swimming speeds and tribution during daylight hours, with larger search volumes and increased conspicuousness of larvae (postflexion and larger) occurring in high larger larvae (Hunter 1981). It is possible that est densities in suprabenthic habitats and increasing encounter rates between a suite of smaller larvae occurring higher in the water col predators and larger white seabass larvae in na umn (Love et aI. 1984; Jahn and Lavenberg tural systems would offset the observed steady 1986). The suprabenthic distribution of larger increase in detection and escape responses ob larvae has been characterized as a possible adap served in larger larvae in laboratory trials. This tation to high concentrations of food, for preda could occw' until larvae became invulnerable to tor-avoidance or for maintenance of position on attacks. My laboratory data provide some hint of the continental shelf. However, recent studies of this pattern in the slightly increased predation white croaker, Genyo·menu.s linecLfus, larvae in rates on larvae in the 4.5--7.5 mm size range (see dicate that the suprabenthic distribution of older Figures 2A, 3A). However, white seabass lar sciaenid larvae is probably not related to feeding vae are relatively inactive and become increas (Jahn et aI. 1988). ingly demersal during ontogeny, thus, they My results indicate that this ontogenetic shift might not be subject to significantly higher deeper into the water column by older white encounter rates with predators. This remains seabass larvae (and other sciaenids) may be re speculative, however, and is an area for future lated to their predator-avoidance capabilities. investigation. The dominant planktivorous species encountered in midwater habitats of nearshore southern Cali Implications for White Seabass Early fornia waters are fast-swimming, shoaling Life History pelagics such as northern anchovy, sardine, and Pacific mackerel, Scom.ber japonkus. Potential Compared with white seabass larvae, Cali planktivores in near-bottom habitats include fornia sardine, Sardinaps sagaa~, and northern sciaenids, gobiids, embiotocids, clinids, ser anchovy larvae (co-occurring pelagic larvae in ranids, and various flatfishes (Eschmeyer et aI. nearshore southern California waters) appear 1983). All of these demersal species exhibit some better able to detect and avoid attacks by adult type of ambush, hovering, discontinuous, or northern anchovy predators at comparable close-range mode of predatory behavior, similar stages of larval development (Folkvord and to the attack behavior ofjuvenile white seabass, Hunter 1986; Butler and Pickett 1988). These and all attack at slower speeds than the shoaling two clupeoid species also exhibit schooling be pelagics. Based on my experimental evidence, it havior in the later larval stages. Many species is likely that, as they drop out of the plankton, exhibit a combination of larval adaptations to older white seabass larvae maximize their preda minimize fish predatioJl, including long peliods tor-detection and avoidance capabilities when of transparency (e.g., dover sole, Microsfom:u.s encountering demersal predators. Remaining in paC"ificus; Hunter!), rapid development of avoid the plankton in later larval stages and being ance capabilities (northern anchovy and sardine) exposed to pelagic, shoaling fish predators would and schooling (clupeoids). During early feeding prolong a peliod of extreme vulnerability, while stages, white seabass larvae develop a robust, shifting to a demersal disttibution would place highly pigmented body form and exhibit limited white seabass in habitats to which they are bet mobility. During the early postflexion stage ter suited developmentally. (7-10 mm), white seabass start to abandon a strictly pelagic distribution and become notice ACKNOWLEDGMENTS ably more demersal. By the late larval and early juvenile stage, they are found almost exclusively I would like to thank Refik Orhun. Chris associated with submerged cover (often drift Donohoe, and Donald Kent of the Hubbs Marine algae) or near-bottom habitats (pers. obs.; Allen Research Center in San Diego for providing 2 and Franklin 1988; Orhun ; KramerS). In near shore waters of southern California, other 2R. Orhun. Unpubl. data. Hubbs Marine Research Center, Sea World, San Diego. CA 92109. IJ. R. Hunter. UnpubI. data. Southwest Fisheries 3S. Kramer. UnpubI. data. Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Ser Center La Jolla Laboratory, National Marine Fisheries vice. NOAA, La Jolla, CA 92038. Service. NOAA, La Jolla, CA 92038. 550 MARGULIES: VULNERABILITY AND SENSORY DEVELOPMENT OF WHITE SEABASS white seabass eggs and juveniles. Many staff Bull. 160, 144 p. members of the Southwest Fisheries Center in Folkvord, A., and J. R. Hunter. 1986. Size-specific vulnerability of northern anchovy, La Jolla provided laboratory facilities or helpful Engmtdis mordax. larvae to predation by fishes. advice during the study; they are Roger Leong, Fish. Bull., U.S. 84:859-869. Eric Lynn, Bev Macewicz, Carol Kimbrell, Fridgeirsson, E. Sharon Kramer, Steve Swailes, Mark Draw 1978. Embryonic development of five species of gadoid bridge, Gail Theilacker, and Geoffrey Moser. fishes in Icelandic waters. Rit Fiskideildar 6:1--68. Fuiman, L. A. Henry Orr, Roy Allen, and Ron Keefe provided 1989. Vulnerability of Atlantic herring larvae to preda graphics assistance. John Hunter and two tion by yearling herring. Mar. Ecol. Prog. Ser. anonymous referees reviewed earlier versions of 51:291-299. the manuscript. I am especially grateful to John Groman. D. B. Hunter and the late Dr. Reuben Lasker for pro 1982. Histology of the striped bass. Am. Fish. Soc. Monogr. No.3, 116 p. viding the opportunity to undertake this re Houde, E. D. search project. Financial support was provided 1978. Critical food concentrations for larvae of three by a NOAA postdoctoral fellowship through the species of subtropical marine fishes. Bull. Mar. Sci. Associateships Program of the National Re 28:395-411. search Council. 1987. Fish early life dynamics and recruitment varia bility. Am. Fish. Soc. Symp. 2:17-29. Hunter, J. R. LITERATURE CITED 1981. Feeding ecology and predation of marine fish larvae. In R. Lasker (editor), Marine fish larvae: Allen, L. G., and M. P. Franklin. morphology, ecology, and relation to fisheries, p. 1988. Distribution and abundance of young-of-the-year 33-77. Univ. Wash. Press, Seattle. white seabass, Atractosrion nobilis, in the vicinity of 1984. Inferences regarding predation on the early life Long Beach Harbor, California in 1984-1987. Calif. stages of cod and other fishes. In E. Dahl, D. S. Fish Game 74:245-248. Danielssen, E. Moksness, and P. Solemdal (editors), Bailey, K. M. The propogation of cod Gadtts morhua L.• p. 533 1984. Comparison of laboratory rates of predation on 562. Flodevigen Rapp. 1. five species of marine fish larvae by three planktonic Hunter, J. R.• and K. M. Coyne. invertebrates: effects of larval size on vulnerabil 1982. The onset of schooling in northern anchovy ity. Mar. BioI. (BerI.) 79:303-309. larvae, Engraulis mordax. Rep. Calif. Coop. Bailey, K. M.• and E. D. Houde. Oceanic Fish. Invest. 23:246-251. 1989. Predation on eggs and larvae of marine fishes Jahn, A. E.. D. M. Gadomski, and M. L. Sowby. and the recruitment problem. Adv. Mar. BioI. 1988. On the role of food-seeking in the suprabenthic 25:1--83. habit of larval white croaker, GenyonemttS lineatu8 Blaxter, J. H. S. (Pisces: Sciaenidae). Fish. Bull., U.S. 86:251-262. 1986. Development of sense organs and behavior of Jahn, A. E., and R. J. Lavenberg. teleost larvae with special reference to feeding and 1986. Fine-scale distribution of nearshore, supra predator avoidance. Trans. Am. Fish. Soc. 115:98 benthic fish larvae. Mar. Ecol. Prog. Ser. 114. 31:223-231. Blaxter, J. H. S., J. A. B. Gray, and A. C. G. Best. Love. M. S., G. E. McGowan, W. Westphal, R. J. Laven 1983. Structure and development of the free neuro berg, and L. Martin. masts and lateral line system of the herring. J. mar. 1984. Aspects of the life history and fishery of the bioI. Assoc. U.K. 63:247-260. white croaker, Genyonemus lineatus (Sciaenidael, off Brownell, C. L. California. Fish. Bull., U.S. 82:179--198. 1985. Laboratory analysis of cannibalism by larvae of Margulies, D. the cape anchovy Enqra·ul·is capensis. Trans. Am. In press. Vulnerability of larval white perch, Mo Fish. Soc. 114:512--518. ron.e americana, to fish predation. Environ. BioI. Butler, J. L., and D. Pickett. Fish. 1988. Age-specific vulnerability of Pacific sardine, Sar Matthiessen, L. d·inops sagax, larvae to predation by northern an 1880. Untersuchungen uber den Aplanatismus die chovy, Engraul-is mordax. Fish. Bull., U.S. Periscopie del' Krystallinsen in den Augen del' 86:163-167. Fische. Pfluegers Arch. Gesamte Physiol. Menschen Eaton, R. C., and R. DiDomenico. Tiere 21:287-S07. 1986. Role of the teleost escape response during devel Moser, H. G., D. A. Ambrose. M. S. Busby. J. L. Butler, E. opment. Trans. Am. Fish. Soc. 115:128-142. M. Sandknop, B. Y. Sumida. and E. G. Stevens. Eschmeyer, W. N.• E. S. Herald. and H. Hammann. 1983. Description of early stages of white seabass, 1983. A field guide to Pacific coast fishes of North Atmctoscion nobilis, with notes on distribution. Rep. America. Houghton Mifflin Co., Boston, 336 p. Calif. Coop. Oceanic Fish. Invest. 24:182-193. Feder, H. M., C. H. Turner. and C. Limbaugh. Munro, A. D. 1974. Observations on fishes associated with kelp beds 1984. The ontogeny of the retina and optic tectum in in southern California. Calif. Dep. Fish Game, Fish Aequidens ponalegrensis (Hensel). J. Fish BioI. 551 FISHERY BULLETIN: YOLo 87, NO.3, 1989 24:377~93. SAS Institute. Neave. D. A. 1982. SAS user's guide: statistics. SAS Institute, 1984. The development of visual acuity in larval plaice Inc., Cary, NC, 956 p. (Pleuronectes platessa L.) and turbot (Scaphthalmus Tamura. T. maximus L.l. J. expo mar. BioI. Ecol. 78:167-175. 1957. A study on visual perception in fish, especially on 1986. The development of the lateral line system in resolving power and accommodation. Bull. Jpn. Soc. plaice (Pleuronectes platessal and turbot IScophthal Sci. Fish. 22:536-557. mus m.axillmsl. J. mar. bioI. Assoc. U.K. 66:683-693. Vinyard, G. L.. and W. J. O'Brien. O·Brien. W. J. 1976. Effects of light and turbidity on the reactive 1979. The predator-prey interaction of planktivorous distance of bluegill (Lepomis macrochints). J. Fish. fish and zooplankton. Am. Sci. 67:572-581. Res. Board Can. 33:2845--2849. O'Connell, C. P. Vojkovich. M., and R. J. Reed. 1981. Development of organ systems in the northem 1983. White seabass, Atra.ctoscion nobilis. in Cali anchovy, Engraulismordax, and other teleosts. Am. fomia-Mexican waters: status of the fishery. Rep. ZooI. 21:429-446. Calif. Coop. Oceanic Fish. Invest. 24:79-83. Rice. J. A., L. B. Crowder. and F. P. Binkowski. Webb. P. W. 1987. Evaluating potential sources of mortality for 1976. The effect ofsize on the fast-start performance of larval bloater (Co1-egonus hoyi): starvation and vuner rainbow trout. Salmo gairdneri, and a consideration ability to predation. Can. J. Fish. Aquat. Sci. of piscivorous predator-prey interactions. J. Exp. 44:467-472. BioI. 65:157-177. Ricker. W. E. 1981. Responses of northern anchovy, Engra.ulis 1975. Computation and interpretation of biological mordax, larvae to predation by a biting planktivore. statistics of fish populations. Fish. Res. Board Can.• Amphiprion pe1'Cu.la. Fish. BulL •. U.S. 79:727-735. Bull. 191, 382 p. 552