The data sample is 737 calls from October 11, 2001, scanned by hand for "B-only" calls. The scanned time is accurate to only about one pixel (2 seconds), while the "fitted time" determined from the center of the envelope of the filtered call should be more accurate.

Time interval between calls. Most time intervals are seen to be near 50 seconds. A distribution of values going up to 250 seconds or so represents the "breathing intervals" between call sequences.

Time interval between calls, expanded scale. This plot shows the interval calculated from the scan time. The difference is always a multiple of the time per pixel on the spectrograms which were scanned. The cluster of short times near 18:40 comes from a period where there were multiple calling whales whose calls could not be separated.

Time interval between calls, expanded scale, using the "fitted time," based on the centroid of the signal envelope. The distribution appears narrowest near 14:45, where the calls were loudest, giving the best time resolution.

Time interval between calls, scale expanded even more.

Time interval between calls, only for intervals interior to a calling sequence (i.e., not including the "breathing" intervals between two sequences.

Same as last graph, expanded scale. There are no apparent differences from the graphs with all events.

Intensity of the third harmonic versus time interval. There is no great correlation of the time interval distribution with call intensity, except that the very loudest calls may be slightly closer together, and with the interval a bit better defined. The average value and standard deviation for the calls will be treated later.

Call number within the sequence versus time interval. There is no evidence for the call spacing changing during the call sequence.

IDL program rdfit.pro used to make the graphs above.
IDL program pbbdt1.pro used to make the graphs above.