178 EARL DE GREY'S ADDRESS. [May 28, 1860. 



being at its outlet 150 yards broad, 10 to 12 feet deep, and nmning 

 at 2^ knots an hour. Lastly, there is this farther unexplained 

 peculiarity, that, contrary to the Zambesi, and to the properties of 

 all rivers in Tropical Africa, the variation in the height of the 

 Shire in the wet and dry seasons does not exceed the remarkably 

 small amount of 2 or 3 feet. 



Now if we venture to disregard native testimony altogether on 

 that one point in which native testimony is perpetually misleading 

 travellers, namely, the direction of the current of a river, the facts 

 at present before us appear not only not contradictory, but even lend 

 considerable probability to the theory that the Xyassa is connected 

 with the Tanganyika, and that the Shire may be the outlet of both 

 of them, and also to the surplus waters of the Shirwa. 



First, as to the elevation above the sea of the water-levels of 

 these lakes. Speke places the Tanganyika at 1844 feet above the 

 sea. Livingstone places the Shirwa at 2000. He has not yet given 

 us the ■ altitude of the Nyassa, but he reports that its waters are 

 described as being separated from those of the Shirwa by a mere 

 spit of land, which assuredly would be flooded in some seasons (if 

 the Shirwa had no kind of outlet), and a water-way worn between 

 the two lakes if there were not a free intercommunication between 

 them through a porous soil, if by no more direct channel. In this 

 way the surplus waters of the Shirwa might find an ultimate outlet 

 by the Shire. 



Next, as to the recorded depression of 166 feet of the Tanganyika 

 below the Shirwa or the Nyassa. This quantity is far too minute 

 to be relied on as accurate, considering the nature of the observa- 

 tions employed by the two travellers, which were simply the record 

 of the temperatures of boiling water, corrected for the temperature 

 of the air. This simple and excellent method of determining 

 heights approximately is wholly unreliable in a case like this 

 unless special precautions be taken, and certain comparisons be 

 made which have not been made in the present instances. For 

 example, the thermometers require to be. verified at the close of a 

 journey as well as at its commencement, because their index errors 

 are found to vary continually by a slight but accumulative change. 

 Speke's thermometer had varied 1° Fahr. from first to last, which 

 represents an altitude of 535 feet. Again, the variation of baro- 

 metric pressure, though small between the tropics at sea, is even 

 there sufficient to cause an error of 100 feet in any one observation, 

 or a differential error of 200 feet between two observations, sup- 

 posing the variation to have acted in opposite directions ; and the 



