STATISTICS IN CLASSIFYING RACES OF SHAD 



271 



example, 70 percent Hudson shad, 30 percent 

 Connecticut shad) for each stratum. From these 

 quantities, the entire pound-net catch could be 

 divided into two parts, Hudson River shad and 

 Connecticut River shad. 



Obtaining the total catch and effort statistics 

 is rather straightforward and not expensive, but 

 estimating the compositions of the catches is 

 much more complex. At the present time, this 

 would have to be done by tagging experiments. 

 In each stratum, a number of shad would be 

 tagged, preferably at various times throughout the 

 season. If the fishermen are personally inter- 

 viewed to obtain tag returns, the coverage on the 

 two rivers should be equal. Even the fishing 

 rates for the two rivers should be the same or 

 some adjustment would be needed to place them 

 on the same level. As a result, the problem of 

 estimating the composition of a pound-net catch 

 becomes complex and expensive when two river 

 systems are canvassed for tag returns, and the 

 tagging is done several hundred miles from the 

 rivers. 



This is one of many fishery problems where it 

 would be advantageous to obtain a sample of fish 

 and classify them according to the river system 

 to which they belong. If this were done accu- 

 rately, samples of fish from various strata could 

 be obtained, and each fish could be assigned to 

 the proper river. The composition of this sample 

 would be used to estimate the composition of the 

 stratum. The remainder of this paper will in- 

 vestigate the statistical techniques applicable to 

 this problem. 



SOURCE OF MERISTIC DATA 



Races (populations) of fish can often be sepa- 

 rated through the use of body measurements or 

 meristic counts. If some of these counts or 

 measurements are sufficiently different for two 

 populations, it is possible to classify the indi- 

 viduals in a sample and estimate the relative 

 abundance of each population in an area by the 

 composition of the sample. 



A large number of morphological data were 

 collected on the American shad, both juvenile and 

 adults, by the Fish and Wildlife Service between 

 1939 and 1945. Data on twenty-five different 

 characters were collected from each fish. They 

 were defined as follows: 



MID-CAUDAL LENGTH.— Tip of snout to end of 



shortest rays between lobes of caudal fin. 

 TOTAL LENGTH.— Tip of snout to end of longest ray of 



caudal fin. 

 STANDARD LENGTH.— Tip of snout to branching of 



urostyle (modified vertebra). 

 DEPTH. — Longest measurement from dorsal to ventral 



profiles (in front of dorsal fin). 

 THICKNESS OF FISH.— Measurement from left to right 



through thickest part of fish. 

 CAUDAL PEDUNCLE.— -Shortest dorsoventral meas- 

 urement of tail anterior to caudal fin. 

 HEAD LENGTH. — Tip of snout to posterior margin of 



opercular bone. 

 SNOUT. — Tip of snout to anterior margin of eye socket. 

 EYE. — -From anterior to posterior margin of eye socket. 

 INTERORBITAL. — Across top of head from dorsal 

 margin of one eye socket to dorsal margin of the other 

 eye socket above pupil of the eye. 

 MAXILLARY. — From posterior margin of maxillary to a 



vertical from tip of snout. 

 LENGTH OF DORSAL AND ANAL BASES.— From 

 anterior margin of base of first ray to posterior margin of 

 last ray. 

 LENGTH OF PECTORAL.— From articulation of first 



ray to tip of longest ray. 

 SNOUT TO DORSAL. — Tip of snout to articulation of 



first ray of dorsal fin. 

 SNOUT TO ANAL. — Tip of snout to articulation of first 



ray of anal fin. 

 PECTORAL TO VENTRAL.— From articulation of first 

 ray of pectoral fin to articulation of first ray of ventral 

 fin. 

 VENTRAL TO ANAL. — From articulation of first ray of 



ventral fin to articulation of first ray of anal fin. 

 ANTERIOR SCUTES.— All scutes having processes in 

 front of ventral fins, including the scute between the 

 fins which does not appear to have a process. It is 

 beneath the process of the preceding scute. 

 POSTERIOR SCUTES.— All scutes posterior to ventral 



fins. 

 VERTEBRAE. — Urostyle included in count. 

 DORSAL RAYS.— Last undivided ray counted with 



divided rays, other undivided rays separate. 

 ANAL RAYS.— As for dorsal rays. 

 PECTORAL RAYS.— All rays on left and right sides of 



fish. 

 GILL RAKERS.— Only those on the lower limb of the 

 first gill arch counted (at the bend of the arch, the bases 

 of the rakers of the upper arm point in the opposite 

 direction from those of the lower arm). 

 SCALES.— Oblique rows from the upper end of opercular 

 slit to base of caudal fin. Horizontal rows from the 

 median dorsal line to ventral scutes but not including 

 either. 



Since there were numerous rivers involved and 

 samples were taken for several years from some 

 rivers, analysis using all the data would become 

 exceedingly complex. Table 2 gives the location, 

 year, and number of adult specimens examined. 



