Fishes of the Western North Atlantic 317 



level of upper margin of gill cover. Sides light green, with strong tinge of gold, espe- 

 cially on head, shading into the white of belly. No dark shoulder spot. Tips of jaws 

 blackish. Dorsal fin yellowish, more or less dusky at base; caudal yellowish, tips of lobes 

 dusky; anal and pelvic pale; pectoral pale, a dusky streak on inner side behind first 

 ray. Peritoneum pale. 



Size. Skipjacks are smaller in size than Alabama shad, with which they are more 

 or less associated in the rivers; their length usually ranges between about 300-450 mm 

 (12-18 in.). 



Development. Although ovarian eggs 0.08-1. 11 mm in diameter were observed, 

 apparently no young under 107 mm (about 4.33 in.) were secured during Dr. Coker's 

 extensive investigations at Keokuk, Iowa (26; 27). Each of the smallest specimens at 

 hand (UMMZ 122510, 128570), from the mouth of Elk River, Wheeler Reservoir, 

 Alabama, and from the Mississippi River, Miles County, Arkansas, is about 40 mm 

 (3 1 mm SL), apparently the smallest known. They differ from larger examples princi- 

 pally in being more slender, the depth of each being contained 4.75 times in SL. 

 Furthermore, they have minute teeth on the margin of the maxillary, these persisting 

 in examples up to 1 50 mm or so in length. The gill rakers, unlike those in some other 

 species of the genus Pomolobus., do not increase in number with age and growth. 



Spawning and Migrations. The spawning place and length of spawning are not 

 fully known. For such information as is available we are indebted to Robert E. Coker, 

 who, with assistants, carried on an extensive investigation at Keokuk, Iowa (26: 25, 26; 

 27: 94, 165-169). They reported fish with large roe taken there on: April 29 and 

 May 23 in 1914; April 29 and 30, May 24, June 5 and 23, July i and 2, and 

 "various other dates" in 191 5; and on May 23, June 3 and 17 in 1916. Fully mature 

 fish, ripe for stripping, apparently were not included. Coker concluded his remarks on 

 spawning as follows: 



Many attempts made during several years to find a particular place and time where and when river herring 

 were clearly spawning met with no success, although various methods were tried. The problem proved unexpec- 

 tedly baffling. Our data suggest only that the spawning season is prolonged, possibly beginning early in May; 

 that the fish do not spawn in large aggregations; that during spawning operations they are not readily captured 

 by ordinary methods of fishery; and that spawning is ended soon after the first of July. 



It has not been proven that Skipjack are anadromous like most of the near rela- 

 tives, even though Jordan said categorically, ". . . migrating from the Gulf of Mexico" 

 (ji: 40). However, they do enter salt and brackish water. Such a great "run" as was 

 reported at Keokuk, Iowa, in the spring of 19 14 as well as smaller runs in succes- 

 sive years and their apparent disappearance in September of each year do suggest 

 extensive migration. However, Coker expressed the opinion that the migrations do not 

 necessarily extend to extreme northern or extreme southern waters. This is in contrast 

 to the strictly anadromous relatives, Alosa sapidissima, Pomolobus aestivalis, and P.pseudo- 

 harengus (except where the last species is landlocked), whose young leave the fresh- 

 water streams at the end of their first summer and do not return until they are ready 

 to spawn. 



