CHEMORECEPTION: LOCOMOTION AND ORIENTATION 311 



Table 1. Differences resulting from flow and nonflow conditions of six 

 locomotor variables in quadrants 1 (upstream) and 4 (downstream), as the 

 shark responds to a single release site.* 



In perfect concordance with the results of the time series and regression 

 analyses, the precision of localization remains high for release sites 2 and 3 

 but decreases rapidly as the sites move upstream (4, 5, 6), although general 

 localization of the area (quadrant 1) remains good. The possibility that these 

 differences are due to habituation or adaptation can be ruled out, because 

 precise localization of sites 1, 2, and 3 continues for the whole period of 

 stimulation (100 min in the plots shown). The decreasing precision of local- 

 ization of sites 4 to 6 must be attributed to the gradual decrease in spatial 

 definition of the gradient as the chemical mixture moves downstream from 

 the release site. 



Behavior of the shark in stagnant water is strikingly different. Only ex- 

 ceptionally does it hover near the release site, and then only for a short time. 

 Whereas the general area of release is localized by the animal, its responses 

 are not discernibly related to the exact locale of the chemical source. This 

 lack of precise localization is similar to that observed for sites 4, 5, and 6 in 

 flowing water, and the cause may be the same, namely that such localization 

 depends on a steep gradient. In a stagnant medium, the shrimp extract, on 

 leaving the release needle, disperses and undergoes rapid dilution in all direc- 

 tions, resulting in a weak gradient. With flow (which is nearly laminar in the 

 monitor tank), a downstream "corridor" of extract, with a strong mean 

 direction vector, is formed. The conclusion to be drawn from these experi- 

 ments is that "gradient search" is a very unlikely orientation mechanism in 

 the natural environment, where gradients commonly, are extremely weak. 



A FORECASTING TECHNIQUE TO DETECT SUBTLE 

 RESPONSES TO CHEMICAL STIMULATION 



The experiments summarized above have demonstrated that fish frequently 

 respond to environmental stimuli with subtle modifications of locomotor 

 variables rather than with overt changes in swimming direction. Such modifi- 

 cations escape detection by direct observation not only because they are 



