being able to form three small houses of the materials of the second observatory; for there was not sufficient level space to have put it up as one house.
By the twenty-fifth the instruments were all fixed Nov. 25. and adjusted, and we had the satisfaction of finding, during three days' preliminary observations, that the foundation remained perfectly steady, and the results were most satisfactory.
The term-day observations were made on the twenty-eighth, and afforded, as we afterwards found, a most interesting comparison with those made at the Rossbank observatory, Van Diemen's Land, showing the same instantaneous movements of the instruments as occur in the northern regions; and thus our principal purpose of coming here was fulfilled to our wishes.
Hourly and additional observations agreed upon before we sailed from Hobart Town were continued until we had obtained seven days of uninterrupted results, when we considered the magnetometric operations complete: the absolute determinations were next to be attended to; but in these we found very considerable difficulty. The place proved to be a most remarkable corroboration of what I have already said respecting the uncertainty and inaccuracy of magnetic observations made on land. In our course from Van Diemen's Land we found a gradual increase of dip, in exact proportion to the distance we sailed during each day towards these islands, from which we could determine with very great accuracy the amount of dip due to their geo-
