290 CHEMICAL SENSES 



been ignored by most workers. The reasons for this are equally obvious; 

 accounting for variability requires that it be statistically defined, which in 

 turn requires quantitative description of locomotor behavior over time, a 

 task of considerable experimental and theoretical complexity. However, an 

 experimental, quantitative assessment of oriented as well as general loco- 

 motor responses to chemical and/or physical cues in the environment cannot 

 be rationally attempted before spontaneous locomotor behavior and its vari- 

 ability over time, in the absence of such cues, is understood, at least to a 

 certain degree. It is mainly this consideration that gives the rationale for the 

 experimental approaches to the study of orientation in fishes in this labora- 

 tory! 



It was considered that, ideally, the definition of the general locomotor 

 behavior of a species should take the form of a model so that changes in that 

 behavior, resulting from response(s) to sensory cues, might be detected in 

 modifications of that model and, therefore, subject to quantification. 



It became necessary to define those characteristics whose values might 

 adequately describe locomotor behavior over time and, therefore, should be 

 incorporated in the model. A locomotor pattern is the consequence of tem- 

 poral relationships between the magnitudes (direction) and frequencies of 

 turns, the lengths and orientations of the straight pathways or "steps" be- 

 tween these turns, and the velocities of the fish. Experimentally, therefore, 

 the problem was to devise ways to monitor the above variables in free- 

 swimming fish. Cinematography, either continuous or intermittent, quickly 

 proved inadequate for the purpose because variability of the values moni- 

 tored and the need to define that variability almost always required monitor- 

 ing of the movements over extended periods. Furthermore, visual measure- 

 ment of large numbers of angles, steps, and time lapses proved to be 

 impracticable. Nevertheless, some very useful data were gathered cinema- 

 tographically; they lead to the recognition of logarithmic spiral configura- 

 tions in the swimming patterns of Ginglymostoma (see below). 



Consequently, through various stages of improvement, two monitor sys- 

 tems were developed (Kleerekoper 1967, 1969; Kleerekoper et al. 1969). An 

 updated description was prepared recently (Kleerekoper 1977) and is here 

 summarized. The two systems have different but complementary capabilities 

 and are frequently used in conjunction. The first system (monitor I) has 

 three versions in operation in this laboratory. One is for larger fish, such as 

 juvenile sharks up to about 80 or 90 cm long, and two are for smaller fish; 

 the latter monitors, in addition, are equipped for orientation studies involv- 

 ing polarized light. 



Only the shark monitor will be described here. It consists of a cylindrical 

 steel tank, 549 cm in diameter and 103 cm deep (Figure 13). Sixteen radially 

 oriented, hollow boxes (A), 97 cm high and 138 cm long, partially divide 

 the tank into 16 peripheral compartments, which communicate centrally 

 with the remaining open space (241 cm) of the tank. Near their central, free 

 extremities, the lateral walls of the boxes have narrow vertical windows. In 

 alternate boxes, alined behind these windows, are white tubular fluorescent 

 lamps (40W), covered with red filters. Behind the windows of the remaining 



