208 Permanence of the Midlierfurd Photographic Plates. 



330, and in the present series of observations the thread was bi- 

 sected at intervals of five millimetres throughout its length. The 

 following table contains the readings on the thread obtained by 

 two independent observers, Mrs. Herman S. Davis and Mrs. 

 Annie Maclear Jacob3^ Each number is the mean of at least six 

 observations, made in three independent series. In each such 

 series the observations were always repeated in the reverse order, 

 so as to eliminate the effects of an}'' possible continuous changes 

 of temperature. 



Table of Spider Thread Observations. 



Observer Davis. 



Observer Jacoby. 





Spider Line. 



cale. Dir't. Eev'd. 



325 9 



R 1 



4987 9 



4977 



320 



4985 



4970 



315 



4993 



4977 



310 



5000 



4970 



305 



4999 



4967 



300 



4996 



4968 



295 



4997 



4968 



290 



4987 



4950 



285 



5004 



4965 



280 



4980 



4953 



275 



4993 



4965 



270 



4996 



4947 



265 



4995 



4953 



260 



4997 



4955 



255 



4987 



4952 



250 



4991 



4955 



245 



4987 



4947 



240 



4985 



4957 



235 



4966 



4943 



230 



4971 



4932 



225 



4967 



4937 



220 



4962 



4928 



215 



4972 



4937 



210 



4962 



4930 



205 



4964 



4933 



Spider Line. 



Dir't. Eev'd 



K ] 



.5000 9 



5000 



.5007 



4988 



.4997 



4998 



•4998 



4990 



•4997 



4978 



.5017 



4990 



.5002 



4997 



.5000 



4985 



•4983 



4997 



.4988 



4985 



•4993 



4985 



.4980 



4983 



.4987 



4982 



.4982 



4987 



.4977 



4978 



.4982 



4985 



.4987 



4985 



.4968 



4988 



.4972 



4983 



.4958 



4977 



.4962 



4980 



.4957 



4988 



.4958 



4962 



.4940 



4970 



4932 



4968 



If there be appreciable errors of the C3dinder, the}^ should ap- 

 pear from the variation of the numbers in the above table. But 

 first these numbers must be corrected for a possible lack of paral- 

 lelism of the spider thread and the cylinder. It is also evidently 

 allowable to add a constant to all the numbers. The following 

 method of reduction has been used. Let : 



