12
JACK COHEN
present may continue to differentiate; however these often disappear to
be replaced by a new set whose pattern is dependent upon the size and
shape of the expiant. These new "pterylae" often appear to be aligned
to the edges of the expiant, and it is very tempting here to think in
terms of diffusion gradients, thresholds, and "morphogens" (Turing,
1952). Sengel himself has suggested an explanation in these terms, but
until his model shows predictive capabilities it is, perhaps, useful mostly
as a pointer to future work. The development of a predictable point
pattern or grid on a Turing system is possible (e.g., Maynard Smith and
Sondhi, 1961), but without a parameter of experimental variation proves
difficult to assess. Maynard Smith uses genetic differences, and he explains
duplications, etc., in Turing's terms. He cannot, however, make predictions as yet.
Several of the theories of hair-follicle pattern development are relevant here. Those rash enough to propose hypotheses fall into two classes :
those who believe essentially in gradient systems and thresholds (e.g.,
Demeijere, 1894; Carter, 1965) and those who believe in random origins
with areas of inhibition surrounding follicles (e.g., J. H. Claxton, personal communication, 1964). Theories of the first type explain the
development of vibrissa fields perfectly, but require too complex a
system to be credible for the general pelage. On the other hand, the
"random" theories do give patterns resembling those of the general pelage
follicles but fail on the vibrissa pad (e.g., Berry Kindred, personal communication, 1964; Oliver, 1965). Turing's examples do, indeed, provide
a series of possible analogies for both systems. However, until more is
known of the variations in pattern in experimental situations, further
hypothesis is probably useless. In any event it seems certain that there
is some order in pelage follicles, and some latitude in vibrissa or pterylar
pattern; perhaps Turing has the stage for the initial inductions,
which then serve as foci for random "filling-in." Otherwise the initial
appearance of, for example, guard hair follicle rudiments could be
random, but each might then serve to establish a Turing mini-system in
its environs. For descriptions of the detail of pteryla origins, see Hardesty
(1933) and Holmes (1935).
C. The Individual Feather
Some idea of the complexity of a feather should first be acquired by
simple examination. Too few biologists have looked closely at this structure; it is a humbling process. (Perhaps it should be made an obligatory
exercise for the increasing number who refuse to take cognizance of evolutionary levels between Escherichia coli and man. Agassiz is said to have
utilized a similar pedagogic technique.)
JACK COHEN
present may continue to differentiate; however these often disappear to
be replaced by a new set whose pattern is dependent upon the size and
shape of the expiant. These new "pterylae" often appear to be aligned
to the edges of the expiant, and it is very tempting here to think in
terms of diffusion gradients, thresholds, and "morphogens" (Turing,
1952). Sengel himself has suggested an explanation in these terms, but
until his model shows predictive capabilities it is, perhaps, useful mostly
as a pointer to future work. The development of a predictable point
pattern or grid on a Turing system is possible (e.g., Maynard Smith and
Sondhi, 1961), but without a parameter of experimental variation proves
difficult to assess. Maynard Smith uses genetic differences, and he explains
duplications, etc., in Turing's terms. He cannot, however, make predictions as yet.
Several of the theories of hair-follicle pattern development are relevant here. Those rash enough to propose hypotheses fall into two classes :
those who believe essentially in gradient systems and thresholds (e.g.,
Demeijere, 1894; Carter, 1965) and those who believe in random origins
with areas of inhibition surrounding follicles (e.g., J. H. Claxton, personal communication, 1964). Theories of the first type explain the
development of vibrissa fields perfectly, but require too complex a
system to be credible for the general pelage. On the other hand, the
"random" theories do give patterns resembling those of the general pelage
follicles but fail on the vibrissa pad (e.g., Berry Kindred, personal communication, 1964; Oliver, 1965). Turing's examples do, indeed, provide
a series of possible analogies for both systems. However, until more is
known of the variations in pattern in experimental situations, further
hypothesis is probably useless. In any event it seems certain that there
is some order in pelage follicles, and some latitude in vibrissa or pterylar
pattern; perhaps Turing has the stage for the initial inductions,
which then serve as foci for random "filling-in." Otherwise the initial
appearance of, for example, guard hair follicle rudiments could be
random, but each might then serve to establish a Turing mini-system in
its environs. For descriptions of the detail of pteryla origins, see Hardesty
(1933) and Holmes (1935).
C. The Individual Feather
Some idea of the complexity of a feather should first be acquired by
simple examination. Too few biologists have looked closely at this structure; it is a humbling process. (Perhaps it should be made an obligatory
exercise for the increasing number who refuse to take cognizance of evolutionary levels between Escherichia coli and man. Agassiz is said to have
utilized a similar pedagogic technique.)