Blog

If the essential features of a given curve are not readily duplicated in a later experiment—in either the same or another organism—we take this, not as a cue to resort to averages, but as a warning that some relevant condition has still to be discovered and controlled. (p. 104)