APPENDIX V. 363 9 boiling-point observations, computed by Lieut.

Sugden, R.N., the lake-level lies at an elevation of 4366 feet. Speke, on the other hand, gives the lake an altitude of only 3400 feet.

The computation of altitudes from Mr. Felkin's observations while journeying out to Uganda presents great difficulties. The aneroid used ("Newton, No. 8137") has not been verified at Kew. The comparisons with one of Dr. Emin's aneroids (" Naudet ") at Lado, and with Mr. Wilson's " Browning " at Eubaga, though both undoubtedly trustworthy for the locality, and under the pressure at which they were made, furnish conflicting results which we are quite unable to reconcile or to reduce to a common standard.

At Eubaga Mr. Felkin's " Newton " uniformly ranged 1*34 in. higher than the corrected " Browning." If we were to deduct this quantity from Mr.

Felkin's previous readings they would place Magungo (on the Albert Nyanza) about 1100 feet lower than Rubaga, or at an elevation of at least a thousand feet in excess of the truth.

Mr. Felkin's observations at Lado, on the outward journey, exhibit an error of + '03 in., as compared with the mean pressure during the corresponding month, as determined by Mr. Dovyak.* On his * But at Gondokoro, where Mr. Felkin, however, observed only for one day, his aneroid exhibits no error at all ! This shows the