TPWD 1968 F-6-R-15 #1172: Job Completion Report: Fisheries Investigations - Region 5-B, Appraisal of Various Mesh Sizes in Taking Fishes at Lake Corpus Christi, Texas, Project F-6-15, Job D-3

--- Page 1 --- JOB COMPLETION REPORT As required by FEDERAL AID IN FISHERIES RESTORATION ACT TEXAS Federal Aid Project No. F-6-15 FISHERIES INVESTIGATIONS - REGION 5-B Job No. D-3 (3rd of 4 segments) Appraisal of Various Mesh Sizes in Taking Fishes at Lake Corpus Christi, Texas Project Leader: John C. Barron J. R. Singleton Executive Director Parks and Wildlife Department Austin, Texas Marion Toole Eugene A. Walker D-J3J Coordinator Director, Wildlife Services October 17, 1968 --- Page 2 --- ABSTRACT Uniform mesh gill nets varying from 1 to 4 inches square mesh were set monthly at three locations in Lake Corpus Christi. These data were collected and transposed to punched cards for computer analyses. Tentative analysis utilizing the negative binomial distribution provided information on the net mesh sizes to which various fish species were vulner- able. A detailed example analysis with freshwater drum showed that mesh size introduced the greatest variability into the catch distribution, followed closely by location. Except for blue catfish, few game fish were taken in 3-inch and larger mesh. --- Page 3 --- Project No. Fob No. JOB COMPLETION REPORT Texas F-6-15 Name: Fisheries Investigations - Region 5-B D-3 Title: Appraisal of Various Mesh Sizes in Taking Fishes Period Covered: January 1, 1967 to December 31, 1967 Ovjective: To determine the relation between mesh size of gill nets and species -omposition of the catch. Procedures: Monthly netting was conducted at Lake Corpus Christi. The nets used ‘re uniform mesh gill nets 100 feet long and of the following bar measures: ly-, 2-, 2%-, 3-, 3%-, and 4-inches. The nets were set parallel to each ther approximately fifty yards apart. A random design was used to deter- ioe set and run order so that fishing time variation would be minimized. ree stations were established and each mesh size set at each location every A total of 252 settings were made. Station No. 1 was immediately nth ive the dam; Station No. 2 was at Miller's Island, about five miles above ‘he dam; and Station No. 3 in Ramirena Creek, about 11 miles above the dam. "or each fish taken, the following data were recorded: ieail on | OR ee £ Us ° ° Ne OO On Der Species Net mesh size Date collected Station location Body depth (not measured over curvature) Total length Standard length Weight Fishing time (number of minutes net remained in water) Mean water temperature (average of set and run values) Water turbidity (Secchi disc) Mean station depth (average of inner, mid, and outer depth readings) --- Page 4 --- 2. All data were recorded in the field; and upon return to headquarters were typed, checked for accuracy, and mailed to the Data Processing Section at the Austin Headquarters to be transferred to punched cards. This procedure was in preparation for future analysis by electronic data processing equipment. Results: There is substantial literature on the subject of fishing gear and gear selectivity, and the mathematics are sometimes formidable. Beverton and Holt (1957) have covered the subject extensively in their book, although most of their examples deal with marine species and their equations usually presuppose a knowledge of age of the specimen. Since the purpose of this study was not to determine the fishing power or efficiency of gill nets, but instead to determine the composition of their catches (with particular emphasis on game fishes), a simple approach seemed justifiable. Moyle and Lound (1960), Lambou (1963), and numerous other investigators have demonstrated that the negative binomial distri- bution will yield an appropriate unit for comparing catches of the number of fishes of specific species. The negative binomial is a contagious distribution; contagion being used in the sense that if an individual of a species is found in a given area then the probability that a similar individual will be in the same area is increased and implies that fishes are not distributed at random. The negative binomial distribution concerns the number of times a specific number of individuals of a species is taken in a series of samples. The shape of the curve is usually positively skewed indicating that more individual observations occur below the mean than do above it. Generally the zero class (no fish of the species in question taken in a sample) has the greatest frequency. The distribution is described by three parameters: the mean, m; the variance, v; and the coefficient of contagion, K. Com- putation of the first two is common knowledge, and the latter can be approximated by: K = nie v-m Obviously, m = v since division by zero is undefined; therefore, if v =m then the distribution in question cannot be fitted to the negative binomial. The mean, too, is never larger than the variance in the negative binomial. This distribution has the advantage in that transformations are available which tend toward the normal distribution. Moyle and Lound demonstrated that the transformation y; = log (x; + 4K) is more applicable than the easier to use yy =A/ xe oT 4 These transformations are of primary importance if analysis of variance is contemplated. --- Page 5 --- we Table 1 Catch Distribution of the Freshwater Drum Fitted to the Negative Binomial Distribution Number of Calculated Observed Fish per Net Frequency (F) Frequency (f OMNKDUPWNHEO me Ww Nw hd ; FA 6 5 3 3 1 1 1 4 0 0 L 0 1 10) 3 0 2 --- Page 6 --- Table 2 Analysis of Variance of the Catch Distribution of Freshwater Drum Source of Variation Mesh sizes Months Locations Months x location Months x mesh sizes Locations x mesh sizes Second order interaction ; Total * Significant at the 0.5% level. -149016 +475807 - 160780 » 362863 .162521 2819736 - 883419 2014142 32.358169 1.679618 20.580390 1.789221 0.487310 4.401644 0.824874 24.95% aL? --- Page 7 --- Table 3 Catch Distribution Per Mesh Size for Several Species Species Smallmouth buffalo Mesh (in. ) Number of fish per net *Denotes an infinite or negative reciprocal. --- Page 8 --- Table 4 Catch Distribution Per Mesh Size for Several Species Number of fish per net 1 2 Channel catfish Blue catfish Flathead catfish White bass *Denotes an infinite or negative reciprocal. --- Page 9 --- Table 5 Catch Distribution Per Mesh Size for Several Species 1/K 5+ 2 1 Number of fish per net 3 4 0) (in.) Mesh Species 6.47 29 Largemouth bass 35 Warmouth oe eRe keone) oooocoo°o oOo 0O0C00 0 eee eoke) 0 0 ooo°o Bluegill oo0o0o0c°o oooo°c°o oCoO0000 oo0oCoCc oO HoOoooc eo ooo00 ONNOOO°O N oqHoOooo°o HOoOOC 0 3 3 15 *Denotes an infinite or negative reciprocal. Redear sunfish | White crappie --- Page 10 --- Table 6 Catch Distribution Per Mesh Size for Several Species Species , *Denotes an infinite or negative reciprocal. Number of fish per net ll 2 3 4 Black crappie 8 2 0 2 4 4 6 2 6 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 oO _0 0 Freshwater drum 5 4 2 2 3 0 8 4 1 4 4 1 6 2 0 5 0 0 2 om 0 PN | io ooooor | peneees --- Page 11 --- In order to demonstrate the use of the negative binomial a catch distri- bution of freshwater drum was used (Table 1). Some precision was lost due to the combination of all mesh sizes, but’ the difference measured by Chi square is still not highly significant. Since the process is tedious, time did not permit the fitting of any more of the distributions; but a computer program will be prepared during the next segment for this procedure, and all of the species will be tested. An analysis of variance was also conducted with the drum data making use of the logarithmic transformation shown above. This test was intended to show the source of the variation in the catch distribution. Table 2 shows that the greatest variation in catch was attributed to the different mesh sizes with the different locations following closely. The interaction between location and mesh size is also highly significant. The variation attributable to months and monthly interaction was not significant and can be considered sampling deviations. In Tables 3 to 6 the catch distributions of each species taken are stratified by mesh size. Using 36 net sets, a negative binomial distribution was computed for each mesh size. The quantity 1/K was defined by Moyle and Lound as the dispersion index. The greater the value of this reciprocal, the greater the spread of the data. This appears to be true if the distribution fits the negative binomial. In the above mentioned tables, those meshes which show an infinite or negative reciprocal are of little value for capturing a species. Those meshes which show the smallest frequency for the zero-class should be considered the optimum mesh size for the species. Observation of the maximum mesh size for capture shows that nets of 3-inch mesh produce a negligible catch of all game species except blue catfish. It would appear that gill nets of 3-inch mesh or larger will not harm game fish populations which reaffirms present regulations to this effect. During the 16th segment of this project additional analyses will be conducted using the ADP methods. we / i Prepared by John C. Barron Approved by 4 Project Leader Coordinator Date October 17, 1968 Elgin M. C. Dietz Inland Supervisor --- Page 12 --- -10- References Cites Beverton, R. J. H. and S. J. Holt. 1957. On the dynamics of exploited fish populations. Her Majesty's Stationery Office, London. 533. Lambou, Victor W. 1963. Application of distribution pattern of fishes in Lake Bistineau to design of sampling programs. Prog. Fish-Culturist 25(2) :79-87. Moyle, John B. and Richard Lound. 1960. Confidence limits associated with means and medians of series of net catches. Trans. Am. Fish Soc. 89(1) :53-58.