196
M. A B E R C R O M B I E
interest for morphogenesis. It is obvious that long-range chemotaxis
towards or away from some source of diffusible substance could produce
such movement; it is, in fact, all too obvious, since this is the standard
explanation offered for any trend of direction, without consideration of
other possibilities. It will have become clear from the preceding discussion that there is at present no good evidence for the occurrence of
chemotaxis in any kind of animal cell apart from the polymorph,
monocyte and amphibian neural-crest cell.
It is usual to consider that the rather consistent outward trend of
cells such as fibroblasts from an explant in culture takes place under
the influence of a diffusion gradient centred in the explant. This theory
has been most persuasively argued by Ephrussi (1933). Apart from the
lack of data supporting the occurrence of chemotaxis in the cell types
concerned there is, however, evidence against it: the absence of any
effect ascribable to it in confronted cultures (Abercrombie and
Heaysman, 1954), the failure to upset regular outgrowth by standing a
culture on edge and so draining fluid from it (Jacoby, 1936), the lack
of the expected effect of an explant on a nearby isolated group of cells
(Abercrombie and Gitlin, 1964). It is also unnecessary, since the contact
responses known to occur between cells should produce it. Contact
inhibition produces a "selection" of movement towards cell-free space,
and hence predicts radial spread from an explant (Abercrombie and
Heaysman, 1954). It also predicts the healing of a wound made within
a cultured sheet of cells, by extension inwards of the surrounding cells
(found by Wilbur and Chambers, 1942), parallel emigration from an
explant on a cylindrical surface (found by Weiss, 1945), and the
tendency of cultures spreading from an explant on a plane surface to
move so that the culture has a circular periphery (found by Ephrussi,
1933, and others; see Abercrombie, 1961b). Contact inhibition alone
can therefore produce trends of direction. Contact guidance, when
combined with contact inhibition to specify, as noted above, one of the
two preferred directions, can produce a general trend of direction with
high efficiency. Such a trend can be produced from a distance, as
Weiss (1929) showed, since a substrate in the form of a gel, such as a
fibrin clot, can be oriented by tension. Furthermore, the orientation
can be produced by cell populations. This is strikingly and commonly
shown in the "two-centre" effect (Weiss, 1952) in which two explants
placed close together in a plasma clot orientate the fibrin between them
by tension produced as a result of their growth, and correspondingly
become linked by a bridge of cells moving under the influence of contact
guidance. The evidence that growth is responsible rests at present only
on the finding that explants of brain produce no orientation, while
other tissues that, unlike brain, undergo active mitosis, do so (Weiss,
M. A B E R C R O M B I E
interest for morphogenesis. It is obvious that long-range chemotaxis
towards or away from some source of diffusible substance could produce
such movement; it is, in fact, all too obvious, since this is the standard
explanation offered for any trend of direction, without consideration of
other possibilities. It will have become clear from the preceding discussion that there is at present no good evidence for the occurrence of
chemotaxis in any kind of animal cell apart from the polymorph,
monocyte and amphibian neural-crest cell.
It is usual to consider that the rather consistent outward trend of
cells such as fibroblasts from an explant in culture takes place under
the influence of a diffusion gradient centred in the explant. This theory
has been most persuasively argued by Ephrussi (1933). Apart from the
lack of data supporting the occurrence of chemotaxis in the cell types
concerned there is, however, evidence against it: the absence of any
effect ascribable to it in confronted cultures (Abercrombie and
Heaysman, 1954), the failure to upset regular outgrowth by standing a
culture on edge and so draining fluid from it (Jacoby, 1936), the lack
of the expected effect of an explant on a nearby isolated group of cells
(Abercrombie and Gitlin, 1964). It is also unnecessary, since the contact
responses known to occur between cells should produce it. Contact
inhibition produces a "selection" of movement towards cell-free space,
and hence predicts radial spread from an explant (Abercrombie and
Heaysman, 1954). It also predicts the healing of a wound made within
a cultured sheet of cells, by extension inwards of the surrounding cells
(found by Wilbur and Chambers, 1942), parallel emigration from an
explant on a cylindrical surface (found by Weiss, 1945), and the
tendency of cultures spreading from an explant on a plane surface to
move so that the culture has a circular periphery (found by Ephrussi,
1933, and others; see Abercrombie, 1961b). Contact inhibition alone
can therefore produce trends of direction. Contact guidance, when
combined with contact inhibition to specify, as noted above, one of the
two preferred directions, can produce a general trend of direction with
high efficiency. Such a trend can be produced from a distance, as
Weiss (1929) showed, since a substrate in the form of a gel, such as a
fibrin clot, can be oriented by tension. Furthermore, the orientation
can be produced by cell populations. This is strikingly and commonly
shown in the "two-centre" effect (Weiss, 1952) in which two explants
placed close together in a plasma clot orientate the fibrin between them
by tension produced as a result of their growth, and correspondingly
become linked by a bridge of cells moving under the influence of contact
guidance. The evidence that growth is responsible rests at present only
on the finding that explants of brain produce no orientation, while
other tissues that, unlike brain, undergo active mitosis, do so (Weiss,