204 
Greenberg 
that is crucial to generating creative hy- 
potheses (Lorenz, 1971). For example, time 
sampling can often be misleading unless the 
environmental or social context is considered, 
and that is not always possible (Driver, 
1968), As Vowles (1975) points out, a pre- 
occupation with quantification may be an 
understandable response to the subjective 
and introspective approach of early 
psychology. 
The selection of units of behavior must be 
suitable to the questions being asked. The 
hard decisions of “lumping” and “splitting,” 
familiar to taxonomists, must also be made 
when behavior patterns are to be organized 
into manageable groups. This is true at the 
level of individual units in a coordinated 
sequence as well as in more complex be- 
havioral categories. It is often helpful to 
begin with a broad overview of general ac- 
tivity patterns and then subdivide them until 
the appropriate “resolution” is achieved 
(Fig. 1). A neurophysiologist may profitably 
deal with units as fine as a muscle group and 
its associated actions (e.g., the “functional 
unit” of Liem, 1967, cited by Barlow, 1968: 
217), while a population biologist may be 
concerned with a complete communicative 
act and its consequences. 
While too fine a preliminary analysis can 
obscure a behavioral pattern with trivial de- 
tails, if the units of behavior utilized in the 
study of a species are too broad, difficulties 
may emerge that could not have been antici- 
pated until the comparative perspective was 
brought to bear on a problem. For example, 
details of behavior that might suggest a 
physiological mechanism or the evolutionary 
sources of signal behavior can be overlooked. 
This can be a problem in studies of display 
behavior where the individual components 
of the display and/or their time course are 
often not as clearly delineated as the general 
form of the display. A notable exception is 
the comparative study of the vertical “bob- 
bing” movements of displaying males of the 
genus Sceloporus by Purdue and Carpenter 
(1972), Such analysis is useful because the 
body movements in the ubiquitous bobbing 
displays of iguanid lizards have generally 
been treated as a single unit, when they may, 
in fact, represent multiple behavioral pat- 
terns which have coalesced; neck and limb 
movements may represent two alternative 
mechanisms which have become combined to 
effect a display. Some species have several 
bobbing movements for different displays 
which utilize limb and neck movement to 
different degrees or even dispense with one 
type of movement. 
Another difficulty in defining units of be- 
havior is that motor patterns and sequences 
are too easily defined in functional terms 
which combine explanation and description. 
This can interfere with comparative inter- 
pretations and replications of the original 
study. 
The Evolutionary Perspective 
One goal of comparative studies is the de- 
velopment of the evolutionary perspective 
that leads to an appreciation of the genetic 
(and to that extent “innate”) contribution to 
behavior (Brown, 1969). Comparative stud- 
ies are also useful to investigators in sug- 
gesting hypotheses and in analyzing the 
adaptive aspects of natural units of behavior. 
Both analogy and homology are important 
considerations in thinking comparatively 
about behavior. 
In evolutionary biology, Mayr (1969) 
asserts that only homologous characters are 
of taxonomic importance. Homology origi- 
nally referred to an anatomical correspon- 
dence between structures in two animals 
sharing a similar body plan, but, in current 
use, the evolutionary explanation for hom- 
ology has usurped the original meaning of 
the term (Beer, 1974). 
Two characters under consideration by a 
biologist may differ markedly in form or 
function and yet be related through deriva- 
tion from a common ancestral precursor. 
Such relatedness may provide valuable in- 
sights into the causation of dissimilar be- 
havioral patterns and the manner in which 
animals evolve. 
The idea of behavioral homology, however, 
is a continuing source of confusion. The con- 
