IIO BASHFORD DEAN. 



modest rather than excessive). And let the same number of 

 possible variations be present in each of the various stages in the 

 development of the capsule. Accordingly, at the first stage let 

 us grant that variation number 673 happened to be the most 

 favorable one : this then was selected in both the embryo and in 

 the capsular secretion ; at the second step some number, e. g., 

 973 was similarly selected in both ; at the third, number 14 in 

 both ; at the fourth, number 467 in both ; at the fifth, number 

 Soi in both; and at the two hundredth, number 761 in both. 

 In other words, if we reckon the progressive series as extending 

 over but two hundred steps, taking this number merely for the 

 sake of example, as the possibilities are multitude, the chances 

 are obviously remote of establishing the foregoing coincidences 

 in the two series (one representing the development of the egg, 

 the other of the capsule) by constantly repeated selection of for- 

 tuitous and fluctuating variations. For, taking the first coinci- 

 dence of a favorable fortuitous variation of the embryo selected 

 by a favorable fortuitous variation in the capsule, or indeed, even 

 a coincidence of variations which are passable, not optimum, in 

 both, it is evident that we are dealing with an accident of an im- 

 probable character. Thus, for variation 673 of the embryo to 

 have seized upon variation 673 of the capsule there is but one 

 chance in a thousand. Granted, however, that this unexpected 

 result happened : for the second optimum variations in egg and 

 capsule to have coincided, say the number 973 in both, there is 

 again but one chance in a thousand, --or in other words, the 

 chance for coincidence in the favorable selection of the optimum 

 fortuitous variations one and two, the chances would be one in a 

 million ; of variations one, two and three, one in a billion. 



It would be urged, on the other hand, by selectionists that 

 these figures are misleading, since with hosts of variations taking 

 place in each stage of egg and capsule there would always be a 

 coincidence of one favorable (although perhaps not the best) 

 variation in the egg with a correspondingly favorable (perhaps 

 not the best) character in the capsule. But this assumption 

 leads to no essential change in the result, for in this case if the 

 variations do not correspond with fair accuracy, at successive 

 stages the embryo will not fit the conditions of the capsule, and 



