STUDYING THE HABITS OF FISHES. I 1 25 



some changes would then be necessary in the mechanism by which the shutter 

 is operated from the outside of the water-tight box. 



II. ILLUSTRATIVE EXAMPLE OF METHODS OF OBSERVATION. 



BREEDING HABITS OF THE HORNED DACE (SenwtHus atroniaculatus) . 



In the selection of the horned dace, rather than commercial or otherwise 

 more important fishes, to illustrate the present discussion, chiefly two consider- 

 ations have controlled : ( i ) That the matter here presented has not hitherto been 

 published, and (2) that the behavior of the horned dace is so complicated that it 

 affords an excellent illustration of the methods the writer has found successful in 

 the study of the behavior of fishes in the field. The homed dace is, however, 

 not without economic importance. It is sometimes eaten, but its chief value 

 lies in the fact that, more than any other fish in the region in which it occurs, 

 it furnishes bait to the angler. The present account of its breeding habits 

 embodies the results of the observations of many years, or rather of many 

 seasons, but the record is not a final one. It is an outline or sketch, a pre- 

 liminary account from which details are purposely excluded. 



There is no published description of the breeding habits of the horned dace, 

 although its conspicuous nests must often have been observed. Kendall and 

 Goldsborough (1908) publish notes prepared by Superintendent Charles G. 

 Atkins, of the United States Bureau of Fisheries Station at Craig Brook, Maine, 

 on the breeding habits of Semotilus bullaris. They report that they have them- 

 selves seen the nests of this form and give a diagrammatic picture of such a 

 nest with the fish on it. More detailed observations are greatly to be desired, 

 especially since the nest-building behavior described for Semotilus bullaris ap- 

 pears to be intermediate between that described in this paper for Semotilus 

 atromaculatus and that observed by the writer in Hybopsis kentuckiensis and not 

 yet published. 



Observing the fish at nest-building time. — The observer who approaches one 

 of the gravelly brooks of southern Michigan during the latter part of April or in May 

 is likely to have his attention attracted by certain elongated heaps of gravel scat- 

 tered at intervals along the bottom of the stream (fig. 9, pi. cxviii, and fig. 3, text) . 

 These catch the eye, because the stones that compose them are clean, as though 

 scoured, and show their blotches of bright colors. The heaps consequently stand 

 out in sharp contrast to the surrounding bottom, which is everywhere covered 

 with a uniform brown ooze. Each of these heaps has the form of a low, rounded 

 ridge, commonly a foot in width and 2 or 3 inches high, but varying in length 

 from a foot to 16 or 18 feet. The ridges run with the stream, and at the down- 

 stream end of each is an oval pit (P.) 2 or 3 inches deep and as wide as the 

 ridge. Below the pit again is seen a trail of clean sand, which at its begin- 



