In lieu of an abstract, here is a brief excerpt of the content:

19 Who Sits on the Egg of Cuculus Canorus? Not Karl Pearson An early case, applied to the eggs of the cuckoo bird, illustrates literally the feel of substantive as against statistical significance. “The Egg of Cuculus Canorus” (1901) by Oswald Latter was one of the first articles published in Biometrika. Edited closely by Pearson—Porter notes that Pearson’s style of editing stretched the definition of “independent authorship ”—Latter’s article was a thoroughly Pearsonian product. Cuckoos survive by stealing the domestic labor of others. They use other birds to sit on their eggs, sneaking their cuckoo eggs into the nests of the involuntary foster parents. “An explanation is needed,” wrote Latter in an amusing locution, “of the success which attends this imposition.” Professor A. Newton, in his Dictionary of Birds, had offered an explanation. Without attributing any wonderful sagacity to her, it does not seem unlikely that the cuckoo which had once successfully foisted her egg on a reed-wren . . . should again seek for another reed-wren’s . . . nest . . . and that she should continue her practice from one season to another. . . . Such a habit could hardly fail to become hereditary, so that the daughter of a cuckoo which always put her egg into a reed-wren’s. . . nest, would do as did her mother. . . . It can hardly be questioned that the eggs of the daughter would more or less resemble those of her mother. (Newton , quoted in Latter 1901, 165; spelling modernized) The mother cuckoo’s problem resembles the so-called principal-agent problem of industrial organization. Or as Dawkins see it, an “arms race” (Dawkins 1999, 55). How does she succeed? She would succeed if her eggs relevantly matched those of the victim. Egg characteristics (size, color) and propensity to match this variety of cuckoo with that species of victim (reed-wren or whatever) would be 203 selected for. In other words, Newton and Latter were saying, the foster mother can be duped by size or color. The word duped is metaphorical. By 1901 the Mendel-Darwin-and-biometrics debate had neared full pitch, and Pearson, Weldon, and Yule—the biometricians—were central participants in it. But in 1901 no faction or scientist believed “intent” was involved in the cuckoo behavior, merely unit-character inheritance or natural selection or both (Mayr 1982, chaps. 16, 17; cf. Dawkins 1999, chap. 4). The trick for success—indicated by a hatch in the foster mothers’ clutch and the carrying on of that set of genes—is to avoid imposing too obviously on the foster mother, the sitter. But a “too obvious” imposition is something decided by the foster mother, not by an arbitrary limit of standard errors in egg dimensions observed by humans. This is the point that Galton, Pearson, and Fisher entirely missed, right at the origin of the quantitative revolution. An oversized egg would “inconvenience the sitter;” as Latter elegantly put it, and an oddly colored egg would alarm her. Therefore “my [statistical ] enquiry,” Latter explained, has thus resolved itself chiefly into an attempt to ascertain (1) if the eggs of cuckoos deposited in the nests of any one species stand out as a set apart from Cuckoo’s eggs deposited elsewhere; (2) if the same eggs depart from the rest in such a direction to approximate in size to the eggs of that particular species of foster-parent. (1901, 166; italics supplied) He continues—and here is the crux. The method employed is to find the mean (M) length or breadth, as the case may be, thence to compute the standard deviation . . . and then to find 100␴/M, the coefficient of variation. To test whether any deviation is [statistically] significant, Mr is taken as the mean of the whole race of cuckoos and Ms the mean of cuckoos’ eggs found in the nest of any one species of foster-parent: the standard deviation ␴s of such eggs is also ascertained . . . . If the value of Mr ⫺ Ms be not at least 1.5 to 3 times as great as the value of the other expression the difference of Mr ⫺ Ms is not definitely significant. (Latter 1901, 166; italics supplied) Here is what Latter is arguing: he observes that cuckoo eggs are typically larger than other bird eggs in size—in length and breadth. This is prior knowledge and suffices to set the test, in Fisher’s terminology, as one-sided. Cuckoo eggs in the foster family’s clutch tend to be smaller 204 ⱐ The Cult of Statistical Significance...

Share