104 Dip and Variation of the Magnetic Needle. 
potheses on which they are computed, or errors of a different 
kind? They cannot be principally errors of observation. At 
several of the stations where the differences are greatest, such as 
Philadelphia, Columbus, Cincinnati, and Madison, the dip has 
been repeatedly observed by different individuals, and nearly uni- 
form results obtained. Do they indicate errors of the hypotheses: 
on which they are computed? These hypotheses are that the 
isoclinal lines over a limited territory are straight, parallel, and: 
equidistant. Doubtless each of these hypotheses may be erro- 
neous, and if the lines are extended over any very considerable 
portion of the globe, will be palpably so, But how great are 
these errors within the limits of my several tables? The obser- 
vations themselves afford data for estimating the errors, The 
error arising from curvature and want of parallelism of the isocli- 
nal lines, will appear in the: unequal mean differences under the 
several meridians; and that arising from the lines not being equi- 
distant, will appear in the unequal mean differences under the 
several parallels. We have already seen that for each of the 
tables, these errors are much ot than the mean of. sim differences. 
plored. But Prof. Locke cinta for ta sdicabasices an n extraor inary dngre of 
accuracy ; an accuracy much greater than I could claim for my own, and and greater 
than I had seen claimed by any former observer. In his own language he expected 
his ape would be received with ‘ surprise ;’ and I thought it not amiss 
measure neutralize each other by being alternately positive and negative, and 
are still further reduced by division, so that the mean see may be valuable im 
spite of them. T have lately had an opportunity of compar 
observa- 
tions with Prof. Locke’s, by going upon several of his imegliteae The: folio 
is the result of the observations. 
Pror. Locxs. _._-Pror. Loomis. — - 
Dip. . Date. . eR aS Date. Diff, 
1839: Nov, 2, a 75 eataepe 78 SO 
73 40.9 | 1839, Oct. 30. | 73.34.9.| 1841, Sept. 7. 60 
73 20.6 | 1839. Nov. 5. | 73 23.0} 1841. Sept. 6 & 10. | 42.4 
1.9 | 1840. Sept. 18. | -69°52.8 | 1841. Oct. 09 
: 2 | 1859 Bept. 5: |, 69 25 .5 | 1841. Sept. 29, =5.6 
ti, Aug. 26. | 70 27.7} 1841. Oct. 5, —0.2 
€ is, la £0, 1838. April 3. || 713. L-Oct.. 8 12) 
Aine or Asan e places we re the re identically the same 
except arChnelanans ect here I have taken for Prof. Locke's result the mean of a 
year’s monthly observations. My Suak-siidis ao0-vjioicsns encias sas ogee 
ate, panes Sa in aie, "a 
