of observation error are much smaller with eggs hatching 

 simultaneously. The eider (137) lays six to ten eggs before 

 starting its incubating, and all hatch on the same day ; yet, 

 if it were not known that the steady application of heat to 

 the eggs does not begin with this duck until all the eggs 

 are laid, six to ten days could be erroneously added to the 

 period of incubation. 



Another fruitful source of error in computing the days 

 of incubation is brought about by influences which retard 

 or temporarily suspend the embryonic development, i. e., 

 cooling or neglecting the eggs after they have been incubated 

 for a while, a combination often adding several days to the 

 true length of incubation, exemplified by the records of the 

 ostrich and the albatross. It is also to be noted that many 

 observers seem to date the end of incubation when the eggs 

 are "pipped," while others report it as ended when the bird 

 is fully hatched. Many periods are given approximately 

 only, indicating, perhaps, that the observer was unable to 

 keep daily watch of the nest, or did not deem it necessary 

 or important to determine the length of the period with 

 exactitude, this latter possibility being the source of much 

 discord in the records, and giving birth to such statements 

 as "about seventeen days" or "twenty to twenty-eight days" 

 or "after a few weeks the young are hatched" (64). 



The evidence, it seems to me, points very strongly to the 

 existence of a true for specific) incubation period, which 

 under optimum conditions varies little with each species or 

 subspecies; hence, if there be marked discrepancies or dif- 

 ferences in the records of such species and subspecies, it 

 seems reasonable to believe that the records (at least in 

 part) in such cases are inaccurate. The records of hum- 

 mingbirds and wrens are good examples of conflicts, prob- 

 ably to be explained on the score of inaccuracy, or error in 

 estimating the start of incubation, for it seems highly im- 

 probable that there is a difference of four days in the period 

 of the Ruby-throated Hummingbird and the Black-throated 

 Hummingbird, even though they are specifically distinct, 

 while it is more improbable that there is a difference of two 

 days in the incubation of the Carolina and the Florida 

 Wrens, which are but geographical races of the same species, 

 and these remarks apply equally well to the case of the 

 Loggerhead and Migrant Shrikes, and the Western and 

 Eastern Meadowlarks. The Cedar Bird is a good example 

 of how wide a difference can be found in the incubation 

 records of a given species; the internal evidence in this 

 instance convinces me that sixteen days is probably correct. 



There are clear indications that other errors or conflicts 

 have crept in because of typographical (or clerical) mis- 



13 



