THE APPARENT POSITION OF THE ZODIACAL LIGHT. 143 



The first and second groups of remainders from the morning observations 

 agree so nearly with the first and third of those obtained from the evening 

 observations that there seems to be no reason, so far as these groups are con- 

 cerned, for adopting different systems of corrections in the two cases. The third 

 group of morning observations approximately agrees with the fourth group of 

 evening observations in the mean result for latitude, but not in that for absorp- 

 tion. This might suggest the use of different systems of corrections for the two 

 series of observations, if the large average deviation in latitude, 6°.2, from the 

 group under consideration did not show that very little relative weight can be 

 attributed to the corresponding mean ; for the average deviation in absorption, 

 belonging to the same group, is not remarkably large. The relative veights of 

 the means in latitude for the final groups of evening and morning observations, 

 if computed by the ordinary rule,^ would be 26 and 1. Under these circum- 

 stances, the discordant mean can hardly be used independently. Out of the five 

 remainders from which it is derived, four are affected by two peculiar groups of 

 observations in Table 11. Each of these groups consists of seven observations, 

 made about the same time. The limiting numbers of the charts for the first 

 group are 110, 116, and for the second 130, 137. An examination of these 

 charts, in the original work of Jones, will show that his observed positions of the 

 eastern zodiacal light on the corresponding dates were really exceptional, and 

 indicate the action of special causes, which cannot be traced at present. In the 

 method of reduction here employed, the second of these abnormal groups of 

 observations accidentally counteracts to some extent the effect of the other; for 

 this reason the result may perhaps be allowed a small weight, instead of being 

 rejected, as it would practically be with a weight of J^. The course adopted 

 was to combine the fourth group of evening and the third of morning observa- 

 tions with the relative weights 40 and 5; to combine with equal weights the 

 first groups of evening and morning observations, as well as the third group of 

 evening and the second of morning observations ; and to employ the second 

 group of evening observations independently. The four sets of mean remainders 

 thus obtained are as follows : — 



» Chauvenet, Spherical and Practical Astronomy, II. 494, 505. 



