to record the commentary of the 

 trainer in the tank with the whale. 

 The 3-hr recording session began 

 about I hour before a feeding period 

 and lasted until the water level, which 

 initially was about I m above the 

 whales back, was too low to make 

 useful underwater recordings. 



■Speclrographic analyses were made 

 in the laboratory with a "Vibralyzer" 

 (Kay Electric Company) to determine 

 frequency vs. time, and a real-time 

 spectrum analyzer (Spectral Dynam- 

 ics SD330) connected to an X-Y re- 

 corder (Hewlett Packard 7035B) to 

 portray the relative amplitude vs. 

 frequency. The waveforms were moni- 

 tored with either the spectrum an- 

 alyzer in the Scope Time mode or an 

 external oscilloscope. All of the 

 sounds described below were recorded 

 from the hydrophone output. 



The whale was very inactive and 

 emitted no sounds until the water was 

 lowered enough for the trainer to 

 stand in the tank and touch her back. 

 None of the sounds could be consist- 

 ently associated with a particular 

 behavior. However, one type, a "me- 

 tallic-sounding pulsed signal." was 

 emitted nearly every time the trainer 

 tapped the whale lightly on the back. 



A low-frequency "growl" or "moan." 

 similar to one type of sound recorded 

 from gray whales off San Diego. 

 Calif., by Cummings et al. (1968). 

 was produced twice during the re- 

 cording session. The principal energy 

 of this signal recorded from the cap- 

 tive animal was in a band from 100 

 to 200 Hz, with a secondary peak 

 around 1.5 kHz (Figure lA). The 

 duration of the sound was just over 1 

 sec. There was no obvious movement 

 of the blowholes or expulsion of air 

 associated with this vocalization. 



The most common sound was the 

 "metallic-sounding pulsed signal" 

 which consisted of 8 to 14 pulses in 

 bursts lasting up to 2 sec (Figure IB). 

 The pulses had sharp fronts (fast 

 rise times) with energy extending from 

 below 100 Hz to over 10 kHz. and 

 several resonant peaks, the strongest 

 being at 1.4 kHz. This sound occurred 



as often as five times a minute, even 

 when not incited by the trainer. Only 

 occasionally did it appear to be cor- 

 related with exhalation and move- 

 ment of the blowholes. 



Three times during the recording 

 session, a short (0.2 sec), broadband, 

 "gruntlike" sound (Figure IC) was 

 emitted, without movement of the 

 blowholes. Its peak energy was cen- 

 tered at 200-400 Hz and 1 .6 kHz. 



Figure ID shows the underwater 

 sounds of an exhalation followed by 

 a low-pitched, "blowhole rumble." 

 This combination occurred several 

 times. 



Twice, a long "metallic-sounding 

 pulse train" with a repetition rate of 

 about 14 pulses/sec merged into a 

 long, low-frequency "groan" after 

 about 1.5 sec (Figure IE). Except for 

 the much faster pulse repetition rale, 

 the first part of this vocalization was 

 similar to the sound shown in Figure 

 IB. 



Numerous other sounds produced 

 by Gigi during the 3-hr recording 

 session essentially were variations of 

 one of the five types discussed above. 



SOUNDS RECORDED DURING 

 RELEASE OF GIGI 



Unfortunately, we did not record 

 again in the presence of Gigi until 

 she was released on 13 March 1972. 

 The recording and analysis system 

 used for these data was the same as 

 used at Sea World. Shortly after Gigi 

 was lowered into the water from the 

 barge that carried her out to sea, 

 long trains of "clicks" were heard. 

 Although at the time there was no 

 wa> to determine if these sounds, 

 which were unlike any recorded from 

 Gigi at Sea World, actually came 

 from the whale or from another un- 

 seen biological source in the area, we 

 now believe they were emitted by Gigi. 

 The clicks were nearly identical to the 

 clicks we have recently recorded in 

 the presence of gra\ whales in Wick- 

 aninnish Bay, Vancouver Island. Can- 

 ada. 



The clicks recorded in the presence 

 of Gigi are shown in Figure 2. Their 

 principal energy occupied a band 

 from about 2 to 6 kHz, centered at 



3.4 to 4.0 kHz. Click duration was I 

 to 2 msec. Eight minutes and 15 sec 

 after the whale entered the water 

 most boats in the area shut down their 

 engines for our recording. The first 

 burst of 29 clicks was recorded 6 sec 

 later. Three minutes and 49 sec later 

 the boats started their engines and we 

 had to terminate our final recording 

 of Gigi. During the 3 min and 55 sec 

 of quiet-ship conditions we recorded 

 1.304 clicks. The number of clicks 

 per burst (or train) varied from I to 

 833 and the click repetition rate from 



9.5 to 36.0/sec. The longest click 

 train, containing 833 clicks at an 

 average repetition rate of 19/sec. 

 began about 1 min after the boats had 

 shut down their engines. Although 

 the amplitude of the signals varied 

 with time, we could not correlate 

 signal level with the location of Gigi 

 because the animal was not seen dur- 

 ing the entire time of the recording. 



SOUNDS RECORDED FROM 

 GRAY WHALES OFF 

 VANCOUVER ISLAND 



The system used to record sounds 

 in the presence of gray whales in 

 Wickaninnish Bay on the west coast 

 of Vancouver Island, Canada, con- 

 sisted of a cassette recorder (Sony 

 Model TC-126) and a portable under- 

 water listening set (InterOcean Model 

 9()A Bio-Acustik). The useable fre- 

 quency range of the system was 100 

 Hz to 10 kHz. The hydrophone ar- 

 rangement shown in Figure 3 resulted 

 in good quality recordings with the 

 small boat system. 



Since 1967, as many as seven gray 

 whales have been sighted at one time 

 in Wickaninnish Bay. However, all of 

 the recordings described here were 

 from single whales or pairs. At 1725 

 hr on 10 August 1973. several click 

 trains were recorded from a single 

 feeding gray whale in 10 m of water, 

 1.200 m friim shore. Ver\ little wind 



42 



