|
KARP AND TYUS-COLORADO SQUAWFISH INTERACTION 27
<br />2400 h. A single trial (e.g., 0600 h, 1200 h,
<br />1600 h, 2400 h) was composed of fish position
<br />data taken at 0, 1, 2, 4, 8, and 16 min time
<br />intervals for all species. Data collection began
<br />2 min after the observer entered the laboratory
<br />so that fish presumably had time to resume un-
<br />disturbed activities. An additional 15 sec lag
<br />time was allowed at 1200 h and 2400 h due to
<br />logistics of feeding. This sequence (0600 h-2400
<br />h) was repeated at least six times for each species
<br />over a 9 d period.
<br />Aggression and activity patterns of all species
<br />were evaluated by recording agonistic interac-
<br />tions and movements (passage from 1 cell to
<br />another) of randomly selected fish (i.e., individ-
<br />ual chosen for observation was that fish closest
<br />to some point in the tank taken from a random
<br />number table). These data were collected in
<br />three dimensional space (36 cell matrix) during
<br />2 min periods, and sampling was with replace-
<br />ment. We recorded aggression and activity ob-
<br />servations following collection of the fish posi-
<br />tiondata. Atotal of 33 individuals of each species
<br />in tanks 1-4 was observed for 132 min, 51 P.
<br />lucius and 47 A. melas (tank 5) were observed
<br />for 196 min, and 51 P. lucius and 53 L. cyanellus
<br />(tank 6) were observed for 208 min. Nips, chas-
<br />es, and threats were the only forms of agonistic
<br />behavior observed. We considered nips and
<br />chases to be intentional aggressive acts. Chases
<br />were sustained charges resulting in an obvious
<br />displacement of the recipient. Threats were
<br />considered to be instantaneous non-physical
<br />confrontations which may or may not .have
<br />caused displacement of one or both individuals.
<br />We were not able to detect overt aggression in
<br />"threat" encounters and so determined these
<br />to be the least aggressive of all agonistic acts
<br />observed.
<br />Following completion of the space use, activ-
<br />ity, and agonistic behavior observations, species
<br />were separated into monospecific groups. Each
<br />group was offered 10, 13-25 mm hatchery-
<br />reared razorback suckers, Xyrauchen texanus
<br />(from Dexter National Fish Hatchery, Dexter,
<br />New Mexico), and all acts of aggression and
<br />predation were recorded in the following 60
<br />min period.
<br />Some of the initial fish captures died from
<br />infections of "white-spot disease" Ichthyophthiri-
<br />us mult~liis (D. E. Anderson, pers. comm.), and
<br />all subsequent collections were treated with a
<br />solution of 0.1 ppm malachite green + 25 ppm
<br />formalin immediately following capture. We
<br />held fishes in partitioned aquaria in monospecif-
<br />is groups for at least 2 mo prior to experimen-
<br />tation, and no further disease outbreaks were
<br />noted.
<br />Data analyses.-Fish position data was summed
<br />vertically by layer (upper/middle/lower strat-
<br />ification of the water column) by sample period
<br />(e.g., 0600 h, 1200 h, 1600 h, 2400 h) for all
<br />species. We used a XY contingency table analysis
<br />(Sokal and Rohlf, 1981) to compare day, night,
<br />feeding and nonfeeding distribution patterns,
<br />and to investigate potential shifts in vertical space
<br />use by P. lucius when in the presence of N. lu-
<br />trensis, R. balteatus, and Pimephales promelas. Ag-
<br />onistic behaviors per minute of observation were
<br />tabulated for each species by sampling period.
<br />Goodness of fit Chi-square tests, X2 (Sokal and
<br />Rohlf, 1981), were used to evaluate aggression
<br />between nonfeeding and feeding periods for all
<br />species. We evaluated activity rates (i.e., average
<br />number of total cell changes per minute obser-
<br />vation) and movements (i.e., average number of
<br />different cells used per minute of observation)
<br />using two sample t-tests, Student's t (Sokal and
<br />Rohlf, 1981), for diel and species differences.
<br />Vertical distribution patterns, aggression and
<br />activity levels of Ptychocheilus lucius in tanks 5
<br />and 6 were evaluated within each tank and
<br />ranked for an overall view of species differ-
<br />ences, but were not statistically compared with
<br />the behavior of P. lucius in the C. latipinnis tank
<br />because of differing densities.
<br />RESULTS
<br />Distribution patterns. -Ptychocheilus lucius exhib-
<br />itedsome variability in vertical distribution with
<br />respect to time of day and presence of food, but
<br />generally were dispersed throughout the water
<br />column in all six tanks (Fig. 1). The vertical
<br />distribution of P. lucius in the tank containing
<br />C. latipinnis was significantly different from those
<br />in the N. lutrensis tank (x2 = 20.914, df = 2, P
<br />< 0.05), Pimephales promelas tank (X2 = 9.591,
<br />df = 2, P < 0.05), and R. balteatus tank (X2 =
<br />11.339, df = 2, P < 0.05). Vertical distribution
<br />of Ptychocheilus lucius was most similar to that of
<br />L. cyanellus and R. balteatus (Fig. 2), but all fishes
<br />exhibited a distribution pattern that was signif-
<br />icantly different from that of P. lucius in the C.
<br />latipinnis tank (all Xg > 7.014, df = 2, P < 0.05).
<br />The C. latipinnis, A. melas, N. lutrensis, and Pi-
<br />mephales promelas spent less time in the upper
<br />portion of the tank and exhibited less spatial
<br />overlap with Ptychocheilus lucius.
<br />
|