90 THE CONDOR I VoL.VIII

combining the same records into four centers or clusters of observers, north, south, east and west of the city, and selecting the earliest date for each species in each of the four groups, we find the range of variation reduced to a day or two. [See also Table III.] In this way it can readily be seen that every report, no matter how fragmentary, is of value. It may contain one or two observations not noted at any other of the neighboring stations, while its deficiencies are made up in their records. I do not propose to quote a host of dates and figures, which are uninteresting and hard to follow, but I trust I may have made clear to you the value of combined local records and also the probability that the correct statement of migration at any given point will not be the citation of a single date but by some such state- ment as follows: "Stragglers of a certain species were arriving at Philadelphia from April 23 to 27 and the bulk movement occurred on April 28 and 29." Migration constantly invites us to theorize and generally we find that we are working with very slender chains of evidence. What we need in the future, it seems to me, is more detailed and accurate data; and the plan of organizing large corps of observers at several important centers, as above described, is suggested as a means toward that end. TABLE I. Showing average (4 years) date of arrival at four stations within ten miles of Philadelphia and three stations over twenty miles distant. Station A B C DISTANCE FROM PHILADEI. PttlA 6 nil. 8 mi. o mi. Chtura pela&rica Apr. t9 Apr. '>2 Apr. 23 Piranga erythromelas May 6 May 9 May 8 Seiurus aurocapillus Apr. 30 May 3 May 2 Hylocichla mustelina Apr. 27 May 3 Apr. 3o .ayornis phcrbe Mch. 4 Mch. 2o Mch. 27 D E F G o mi. 2 mi. 22 mi. 33 mi. Apr. 23 Apr. 8 Apr. 2 Apr. 23 May 6 May 9 May 9 May 8 Apr. 28 Apr. 29 Apr. 30 May 4 Apr. 29 May I Apr. 3o Apr. 30 Mch. 8 Mch. 24 Mch. 2o Mch. 27 TABLE II. Showing how first arrival reports from stations about Philadelphia are massed on certain days. Chimney Swift ( Chcelura pelafica).

I903 194 9.5 

April I2 arrived at I station April 5 arr. at 2 sta. April 3 arr. at I sta. o .... 6 " 3 .... x " 2x .... x6" 3 .... 3 " 24 .... 4" Wood Thrush (ylocchla muslelina). 3 4 5 April 3 a. at I sta. April  art. at  s. April 3 arr. at  sta. 3o " "6" 6 .... " 4 .... " 5 .... 4" 9 .... " TABLE II[. Illustrating method of combining data from 2 local observers into several ' ' centers. ' ' Chrtura pelajica, i9o2: April 9 ( sta.), April 2 (), Ap}il 2 (5), April 23 () April 24 (2o)  April 26 (), April 27 (). Grouping the.stations in four sections, "centers," and taking the earliest date for each section, we have: Section A, April 2; B, April 22; C, April 9; I), April 22. To:rostoma rujt, 9o2: April 22 (2), April 23 (I), April 24 (4), April 5 (), April 26 (), April 27 (I), May  (). Grouping them we have: Sect. A, April 22; B, April 22; C, April 23; D, April 24.

Philadelphia, Pa.