emigrants were the same as have been de- 

 scribed earlier for the shrimp moving out of 

 Buttonwood Canal. The shrin-ip during periods 

 of high relative abundance observed in January 

 and September at Joe River were relatively- 

 large juveniles, which averaged 14.0 mm. (0.6 

 inch) carapace length or larger, and apparently 

 entered the catches of small commercial 

 shrimp 1 month later. Shrimp taken during the 

 period of abundance observed in June were 

 small juveniles, which averaged 11.2 mm. (0.4 

 inch) carapace length, and apparently entered 

 the commercial fishery Z months later. 



Progress was made in the computer analysis 

 of the data collected from the channel net in 

 Buttonwood Canal, This analysis is designed 

 to determine the environmental factors most 

 important in influencing fluctuations in abun- 

 dance of the emigrating juveniles and ex- 

 pressing them in quantitative ternns. A multi- 

 variate regression computer program devel- 

 oped by the University of California, Los 

 Angeles, is being used. This program produces 

 a series of multiple linear regression equations 



in a stepwise n-ianner. At each step, the 

 program adds one variable to the regression 

 equation. The variable selected is the one 

 that makes the greatest reduction in the re- 

 maining variation (error sunn of squares). In 

 progressive steps, the residual variation is 

 reduced until all variation is accounted for or 

 until all the variables have been entered and 

 the rennaining variation is reduced to a mini- 

 mum. The end product is a listing of independ- 

 ent variables ranked according to their con- 

 tribution in reducing variation, with their 

 corresponding regression coefficients. The 

 regression coefficients nnay then be used in 

 a prediction nnodel to estimate the abundance 

 of emigrating shrimp. 



C. P. Idyll, E. S. Iversen, and 

 B. J. Yokel, Project Leaders 



Institute of Marine Sciences, 

 University of Miami 

 (Contract No. 14-17-0002-188) 



SHRIMP DYNAMICS PROGRAM 



To attain maximum use of a renewable 

 resource, such as a shrimp population, de- 

 tailed information describing the dynamics of 

 that resource must be obtained. Major aims 

 of our research are to develop reliable tech- 

 niques that can be used to forecast shrinnp 

 abundance before the fishing season, to 

 describe patterns of movement of juvenile 

 and adult shrimp, and to establish realistic 

 estimates of their growth and mortality. 

 Substantial gains were made during the 

 year in the assessment of changes in the 

 shrimp fishery and in the development of 

 sophisticated techniques for predicting the 

 harvest. 



Analysis of systematic collections of post- 

 larval and juvenile brown shrimp indicated 

 that abundance indices developed for these 

 stages can be used to predict the relative 

 magnitude of offshore commercial shrimp 

 supplies. Although variation is greater in 

 collections of postlarvae, predictions based 

 on them have greater potential value than 

 those made from juvenile shrimp catches 

 because the data on postlarvae are available 

 earlier. The use of a plankton pump is being 

 investigated because the variation in post- 

 larval collections may be a result of in- 

 frequent sampling. The punnp will provide 

 samples of migrating postlarvae at frequent 

 intervals or continuously. The advantage 

 gained by continuous sampling is that es- 

 sentially all of the variation caused by en- 

 vironnnental factors can be considered as 

 random. 



Several nnark-recapture experiments were 

 made along the Texas coast to measure growth, 

 movennent, and mortality of brown shrinnp 

 as they ennigrated from estuaries and entered 

 the comnnercial fishery. A small internal 

 plastic tag was used for the first time in these 

 studies. Additional experiments are planned 

 for estimating survival snd growth rates of 

 brown and white shrimp in Texas and 

 Louisiana waters. Such studies provide data 

 required to construct and test mathennatical 

 models of exploited shrimp stocks. Analysis 

 of data from mark-recapture experiments on 

 the Tortugas (Fla.) pink shrimp grounds pro- 

 duced three similar but independent estimates 

 of natural and fishing mortalities. Rates of 

 natural mortality were lower than have been 

 reported previously. 



Four manuscripts were completed during the 

 year on shrimp discarding practices in the 

 Gulf of Mexico, crop prediction from postlarval 

 shrimp indices, the Tortugas shrimp fishery, 

 and estimating shrimp populations. 



Contract research by the USL (University 

 of Southwestern Louisiana) was devoted to the 

 systematic sampling of postlarval and juvenile 

 white and brown shrinnp to deternnine their 

 seasonal abundance in Vernnilion and Cote 

 Blanche Bays. Also, USL made laboratory 

 experinnents to determine if protein extracts 

 from postlarvae would aid in separating the 

 species at that stage. 



Kenneth N. Baxter, Acting Progrann 

 Leader 



14 



