STUDY OF FREE-RANGING SHARKS 431 



stage is off most of the time. This is because the relatively low-drain receiver 

 circuit (on 100% of the time) still has an average drain exceeding that of the 

 power-output stage (which is on a small percent of the time, perhaps 1%). 

 Thus, a net power savings can usually be achieved only for transmitters 

 which are of relatively high average power such as the CSULB shark units. A 

 further means of lowering power drain, as suggested by M. Yerbury (personal 

 communication), would be to impose an on-off duty cycle on the receiver 

 itself, this being controlled by an even lower power clock circuit. For ex- 

 ample, the transponding receiver might be turned on for only 10 ms every 

 100 ms and thus would still be able to respond to any interrogate pulse more 

 than 100 ms in length. 



Besides transponding, other command functions of potential use are (1) 

 interrogation of memory bank (see next section), (2) release of transmitter 

 from shark, and (3) delivery of stimulus to shark for experimental purposes. 

 Recovery of stomach-implanted transmitters might be possible by a com- 

 mand-regurgitate mechanism which releases an emetic. If several different 

 command functions are included in one transmitter, the interrogate signals 

 must be frequency coded for proper recognition. 



Clock-Timed Function (e.g., Timefix)— Some of the advantages of 

 transponding can be achieved without the need for interrogation by using 

 timed functions within the transmitter. One example is battery conservation. 

 If it is not intended to track a shark continuously but only at certain pre- 

 determined hours of day or night, the transmitter can use a micropower 

 clock circuit to turn the transmit section on and off at the desired times. For 

 instance, in a long-term study of home-range stability in a reef shark, it 

 might be necessary to check the shark's position only once a day. The unit 

 could transmit perhaps only 1 h each day, such as between 0800 and 0900 

 each morning, thereby increasing battery life by a factor of nearly 24. 



In regard to distance determination, the "timefix" system developed by 

 the fish-telemetry group at Trondheim, Norway, is of considerable interest 

 (Holand et al., 1974). In timefix USTs, the timing of output pulses is very 

 accurately controlled by a 1-Hz quartz-crystal circuit similar to that in a 

 digital wristwatch. A similar 1-Hz oscillator at the receiver is initially set in 

 exact synchronization with the unit to be applied to the fish. Therefore, any 

 delay in reception of the fish pulse is due to the time needed for that pulse 

 to travel through the water and is thus a measure of distance to the fish. The 

 usefulness of this kind of system, of course, depends greatly on the accuracy 

 (or predictability) of the two oscillators. The longer the time since the initial 

 synchronization, the more uncertain the range information obtained, espe- 

 cially if the temperatures of the two units differ by a substantial but un- 

 known amount. This temperature differential will probably be the main 

 source of error in a timefix system for fish tracking, therefore the tempera- 

 ture coefficients of the crystals become important criteria for their selection. 



Assuming a frequency difference between the two clock oscillators of 10 

 ppm (equivalent to a drift of 1 s/day), a range-accuracy drift of about 



