A total of 313 pigtoes were measirred to detemiine the size 

 composition of the population on the Wheeler beds. Sizes ranged from 

 15 to 105 millimeters, with a mean of 70. 7 mm. Analysis of the length 

 data appears below. 



Number Length range Mean Standard Standard 

 exam-l ned (mm.) (mm.) deviation error 



313 15-105 70.7 12.9 0.8 



The data approach a normal distribution with some negative skewnessj 

 over 60 percent of the samples were larger than the mean length and 

 less than kO percent were smaller than the mean. 



The frequency of occurrence of small- sized mussels on bottoms at 

 the k8 sampling stations was extremely • low . Pigtoes less than 30 mm. in 

 length occurred at only 3 stations. The number of mussels less than 30 

 mm. in size collected from the 96 square yards of bottom area was as 

 follows: 5 pigtoe, 12 deertoe, 3 white wartyback, 2 heelsplitter, 2 

 washboard, and 1 three-horn. 



Relation between abundance and catch . — The resiolts from SCUBA 

 samples in 5^000 square yard test areas in 2 reservoirs showed that 

 population densities varied from 6.7 mussels per square yard in Wheeler 

 to 16.7 per square yard in Chickamauga Reservoir. The population den- 

 sity for the pigtoe was 2.5 per square yard in Wheeler and 10. 7 per 

 sqiiare yard in Chickamauga. 



Brail samples taken in these areas during August and September 

 provided a comparison of catch rates from different populations. In 

 Wheeler Reservoir the mean catches per drag per square yard were 0.02i<- 

 for all species and 0.011 for the pigtoe; in Chickamauga they were O.O56 

 for all species and 0.039 for the pigtoe (table 12). The rates of deple- 

 tion per drag for the different populations amounted to O.36 percent in 

 Wheeler and 0.3^ percent in Chickamauga for all species combined. For 

 the pigtoe, populations were depleted at the rate of O.kk percent per drag 

 in Wheeler and 0.37 percent in the Chickamauga test area. These rates are 

 valid only for the specified sampling periods (August and September) 

 because of seasonal variations in the effectiveness of the harvest method. 



In the treatment of the data for regression analysis, four vari- 

 ables were used to define the correlation between population density 

 and catch success (table 13). The test for correlation resulted in a 

 significant positive coefficient (r= O.98). Using the linear regres- 

 sion method, a line was fitted to the plotted points with the formula 

 Y= -0.001 / O.OO36X (figure 5). The results indicate that fluctuations 

 in catch per drag are directly proportional to changes in the level of 

 abundance during a given time period. 



21 



