Figure 11. Macro-behavioral kinematic organization for the entire 24-hr. period based on all 5 experimental days combined. Connecting arrows have proportional widths depicting the probability of occurrence of each type of behavioral sequence. Numerical data corresponding to these widths are given in parentheses.
Figure 11 summarizes the macro behavioral kinematic organization for the entire 24-hr. period, based on all 5 experimental days combined. Submerged floating had the highest behavioral frequency (accounting for 1/3 of all behaviors occurring) and thus marks the entry into the kinematic graph with a probability of occurrence of .33. From submerged floating, the animal sequenced into head-bob breathing with a .50 probability. Once the animal had moved from submerged floating into head-bob breathing (p = .50), he almost always returned directly to submerged float (p = .92). This relation also obtained for a "submerged float-surface float-submerged float" cycle, although the probabilities were lower (p = .49 for such a closure). The cyclic behavioral variations between submerged floating and breathing most frequently were broken by the animal's moving from surface-float breathing into free swim activity. Once this had occurred, the animal rarely went from free swim directly into submerged floating or head-bob breathing; rather, he was most likely to return to surface-float breathing (p = .79).
Figure 12. Kinematic macro-behavioral organization variations within
the 24-hr. circadian period as depicted by mean probability of each behavioral
sequence across successive 2-hr. periods. The 24-hr. total means are depicted
as the center reference fine in each connecting arrow. Fluctuating arrow widths
reflect the 24-hr. circadian variation in kinematic probabilities.
Figure 12 demonstrates the changes in kinematic behavioral organization across successive 2-hr. blocks for the 24-hr. day. Static forms of kinematic pattern summaries for these same time blocks are given in Figure 13. Four different general systems of behavioral organization are suggested by Figures 12 and 13. The hours of darkness were characterized mainly by the high frequency of submerged floating, with a displacement of subsequent surface-float breathing by head-bob breathing after midnight. During the morning and evening twilight hours, surface-float breathing was most prominent, with a high probability of subsequent free swim activity. By 10:00 a.m., free swim had become the most frequent behavior, with a very high probability of sequencing into surface-floating and then back into free swimming.
Figure 13. Static depictions of kinematic macro-behavioral
organization within successive 2-hr. blocks across the 24-hr. circadian period
for all 5 experimental days combined. The connecting arrows have proportional
widths depicting probability of occurrence of each type of behavioral sequence.
The initial behavior always represents the highest frequency behavior relative
to all others.
By noon, the animal had returned to a submerged floating-head-bob breathing pattern with quite a high degree of behavioral pattern complexity relative to the similar primary sequential organization that occurred during the darkness hours. Thus, the afternoon patterns appear to be a complex composite of the variety of more restricted patterns occurring during other portions of the day. This concept of varying pattern complexity is reflected by actual counts of the total number of different kinds of behavioral sequences occurring within each 2-hr. time period. Such counts represent the total number of connecting arrows in each kinematic diagram. Figure 14 illustrates these sequential pattern variability measures for the 12 2-hr. periods.
Figure
14. The number of macro-behavior sequence pattern variations and the number
of macro-behavior-to-behavior changes (macro- behavioral flow rate) across successive
2-hr. periods within the 24-hr. circadian period.
Pattern variability increases just before noon from a rather consistent a.m. simplicity of 8-11 different sequential occurrences. By mid-afternoon, the complexity has increased by 75-80%, with two periods containing 15 different pattern variations. Few, if any, extremely low sequence probabilities exist (Figures 12 and 13), reflecting a more frequent reliance on even the most tangential patterns compared to the a.m. periods.
The rate of behavioral change (flow rate) within these same 12 2-hr. time blocks also is illustrated in Figure 14. This analysis procedure is analogous to investigating accelerative-decelerative functions (i.e., changes in rate across time). Not only does behavioral patterning variability fluctuate systematically across the day, but also the rate of behavioral change varies. In fact, the two measures are synchronized in time, with increased variability being accompanied by increased rate of flow.
Behavioral Kinetics: Micro Categories
Since the
macro behaviors just discussed verified the behavioral observations of Stage
I, additional micro kinematic pattern analyses were not deemed necessary. However,
we did pursue micro-behavioral sequencing sufficiently to obtain behavioral
flow rate measures for the 9 hr. during which we had taken videotaped recordings.
Behavioral flow rate changes occurring from early morning to late afternoon daylight hours are illustrated in Figure 15. Also depicted in Figure 15 is a continuous illustration of the corresponding macrobehavioral states. These running records of macro-behavioral states are consonant with the a.m. and p.m. kinematic diagrams shown earlier (i.e., Figures 12 and 13).
Figure
15. Successive 30-sec. plots of the total number of micro-behavioral changes/min.
(micro-behavioral flow rate) occurring from early morning until late afternoon
hours. Macro-behavior states corresponding to micro-behavioral flows are illustrated
on the time lines. Magnification insets depict successive 5-sec. plots of micro-behavioral
flow rate changes during bouts of Head-Bob Breathing which occurred between
episodes of Surface-Float Breathing (the most probable type of sequence, p =
.92).
Figure 15 illustrates, in micro detail, the oscillatory character
of behavioral flow rate within the macro- behavioral states. Submerged
floating involved stable periods during which no micro-state changes occurred
(i.e., flow rate was zero). Head-bob breathing, depicted by the specific and
short-lived peaked oscillations, was highly consistent and stereotyped in its
flow rate changes. The magnified inset graphs of Figure 15 illustrate better
how head-bob breathing was marked by regular high frequency oscillations in
micro-behavioral flow rate. These insets are successive
5-sec. summaries (as opposed to the 30-sec. periods of the main figure) illustrating
the periodicity of the "submerging--surfacing--breathing--submerging cycles
which mark the head-bob breathing category.
Free swim activity and surface-float breathing were highly fluid and mutually integrated during this period of the day (see Figures 12 and 13), and thus the two states collectively are represented by relatively long periods of increased behavioral flow rate. Bouts of free swimming and surface-floating were of combined longest total duration (Figure 7) and were most interrelated sequentially (Figures 12 and 13) during the a.m. hours; consequently, these episodes were chosen for a detailed analysis of flow rate oscillations within bouts.
Behavioral flow rate for bouts of combined free swimming plus
surface-float breathing (active behaviors) was averaged across bout duration
for each successive 30-sec. period beginning with the transition from submerged
floating (the most probable base behavior). In addition, the bouts of similar
active behaviors immediately following the three show performances were analyzed
and averaged separately. The analysis makes specific the behavioral perturbations
immediately following the interjections of show periods. The results of these
analyses are depicted in Figure 16.
Figure 16. Number of micro-behavior changes (micro-behavioral flow rate) during each successive 30 sec. immediately preceding and subsequent to behavioral shifts involving (a) Submerged Float-to- "Activity" changes, and (b) Show Performance-to-"Activity" changes. "Activity" is combined Free Swim Activity and Surface-Float Breathing bouts.
Potential relations between oscillatory behavioral flow rate patterns in the rat and killer whale experiments were the subject of another report by the present authors (Ray et al., in press). The primary purpose of that paper was to explore the potentials of formal modeling applied to the comparison of rat and whale data reported above. For example, in the case of data illustrated in Figure 16, Ray et al. assumed the existence of an adaptive control system. Post-show performance was envisioned as representing the system's free response, and the post-submerged floating data as representing this same free response plus the addition of a forced response. A simple model of the post-show data consisted of a constant forward transfer and a single negative feedback transfer function (Brewer, 1974). Where X (s) and Y (s) represent the system's input and output, respectively, in the complex variable domain (s), the canonical form was modeled as:
with Wo and alpha both being constants (Ray et al., in press).
The equation corresponding to this diagram was determined and a least-squares fit between the model equation and the actual post-show performance data was accomplished (see Ray et al. for graphical details). After determining the free response, the steady-state solution for the post-submerged floating data also was derived. A clear characteristic of both data sets was the transient response. Further indication that the data observed may not have been random derived from the fact that a single sinusoid approximated the steady-state data. However, a cosinar analysis (Halberg, Tong & Johnson, 1967) was indeterminant on this issue, probably due to the very small sample size.
The model equation derived for post-submerged float data was then used to probe the rapid pace S- oscillations observed in the rat experiment reported above. The determined equation was:
where Beta depicts the phase angle of the steady-state response. Determined
parameters necessary for modeling the rat rapid pace and the killer whale
post-submerged floating are summarized in Table 1, along with comments on
the implications of these parameter comparisons.
Slow pace rat data required a second-order differential equation with no damping
factor. Because the dominant term determined for the slow pace condition approximated
the rapid pace angular frequency, Ray et al.
(in press) speculated
that driving forces for both pacing conditions had common angular frequencies,
thus implying a common oscillation mechanism. However, cosinar analyses indicated
no periodicity in the rapid pace data, but did confirm the reliability of
the slow pace oscillations.
Clearly, the rat and whale situations being modeled involve many differences which still need to be explored in more detail. But our present purpose is served by the mere demonstration that such newly conceived phenomena exist as the result of a fully integrated systems approach to organisms. We have yet to consider other, no less important, aspects of the interbehavioral system as well. One of these aspects involves the spatial settings of organisms.
Spatial
Factors in Interbehavioral Organization
Whether viewed from
the perspective of macro- or micro-behavioral states, the organism demonstrates
organized behavioral patterning which
can be understood only through spatial as well as temporal analysis. Our studies
clearly demonstrate the temporal rhythmicity of respiratory-behavioral patterning,
but thus far the only data presented relating to the
spatial specificity of
behavioral states involve the water outlet and its role in submerged floating.
An adequate understanding of more subtle environmental controls requires as
complex and detailed analyses as have been accomplished relative to temporal
factors. A hint of the form that such analyses might take and the resultant
insights into behavioral-environmental organization can be illustrated by a
brief example.
For the following analysis, the tank's circumference was divided into 13 equal reference segments. Spatial positions of the animal relative to these segments were determined from videotape recordings made during Day 5 of Experiment 2 for: (a) breath occurrences during surface-float breathing, (b) prolonged submerged positioning during free-swimming bouts, and (c) physical contact points made with the tank while free swimming. The frequency distributions of these measures for the period between 8:00 a.m. and 1:00 p.m. (excluding show performance time) are depicted in Figure 17.
Figure 17. Frequency of occurrence within each specific section of the holding tank of (a) breaths occurring during floating, (b) prolonged submerged positioning in one location during free swimming bouts, and (c) physical contacts made with the tank wall and objects while free swimming. The frequency distributions are for the period between 8 A.M. and I P.M., excluding show performance time, for the last day of observations (Day 5).
The measures for surface-float breathing (Figure 17a) show a clear concentration in the adjacent areas centered around the water outlet, which also include the gate leading into the show tank and a shallow slide-out ledge. Because it has been suggested (Spencer et al., 1967) that sleeping occurs during surface-float configurations like those recorded, our data suggest that the killer whale consistently may sleep within a relatively specific location, as is quite common with most terrestrial mammals.
Prolonged submerged positioning in one location or another during free-swimming periods occurred predominantly at gate areas (Figure l7b). The heavy concentration of these occurrences at the show tank gate is probably accounted for by the fact that two dolphins were swimming in the show tank and were frequently within the vicinity of the gate during such observations.
Physical contacts made with the tank walls during bouts of free swimming (Figure 17c) were much more evenly distributed around the circumference than were the other measures. Although virtually all areas of the tank were contacted, the animal did seem to concentrate somewhat on areas affording corners and ledges (i.e., gate areas), most often used to rub against.
Data depicted in Figure 17 indicate the specificity of behavioral organization of space. If similar analyses are accomplished within successive time periods of shorter duration, as with our analyses of temporally organized behavioral rhythms, more precise variations in composite temporal-spatial-organismic relations may be detailed. As these spatial factors are further incorporated into analyses of temporal boundaries defining relative contextual, discriminatory, and momentary stimulus events, more detailed understanding of individualistic adaptations within both static and changing environments may emerge. Preliminary analyses of such individualistic adaptations to spatial cues, as well as to ambient and punctate discriminatory cues, are already available in reports of laboratory manipulations (Ray, 1977).