C—la

14

The variation of the mean monthly diurnal inequality of D and of H has certainly a well-defined annual part, and the analysis of the curves for a number of years must be undertaken before the connection between diurnal and annual variations can be made clear. I have not had time, to investigate the question whether the seasonal diurnal variations in H in the years 1905, 1910, 191 I, and J9J9 are related at all in the same way as the, annual marches of Hin these years have been shown to be related it seems most probable that they are. ; and, if so, some relationship will be definitely shown to exist for Christchurch between the annual and diurnal variations. It would indeed be pleasant to have a clerical stall' at the Observatory sufficient to enable, such work to be rapidly dealt with. One naturally wants to get the cream from the milk, and such is only possible by the proper machine. As further showing the connection between annual marches of H.F. five years apart, it is notable that the year 1918 of sun-spot maximum was five years after the year of sun-spot minimum, 1913. The curves of annual inarch of H.F. for these years (see diagrams) exhibit a, surprising general resemblance, much more than we should expect if sun-spottcdness was the only variable coi related with the trend of the curve*. It is noteworthy that, the area! differences of these two curves are equal on either side of the mid-year ordinate, suggesting that the excess of the drop from January to April in 1918 was exactly balanced by the excess of the drop from June to September in 1913. The total change of H.F. from January, 1913, to January, 1914, happened to be only ly less than the total change from January, 1918, to January, J919. The mid-points of excess of fall differ by four years and a half, changing in that time from late winter to late summer. The mid-points of defect differ by five years and a half, changing in that time from late summer to late winter. After October they seem to become opposite, the similarity of trend being confined to ten months of the year, a, fact apparently not unconnected with what was found in the mean curve of H.F. for 1911 and 1919, that symmetry could only be found to exist over ten months of the year, and that beyond that the asymmetry exceeded possible observational errors. The curve for the previous sun-spot minimum. 1901, is not available for Christchurch, as magnetograph registration only commenced here at the end of IDOL The sun-spot maximum previous to 1918 was in 1907, and measurements of magnetograms for that year are in progress. Nine, months of declination curves have been measured and declination values got out, but it is not expected to have H.F. hourly values tabulated for 1907 loi some months yet. The measurements of 1912 magnetograms have been completed, but the conversii I mil mates and tabulation arc still incomplete. It is hoped to complete if this year. The provisional value of H.F. for 1912 is 0-22470, which shows a small secular change of JBy from 1911. Similar small changes occurred from 1904 to 1905, and from 1919 to 1920, and were both accompanied by marked change in the form of the curves representing the inarch of H.F. throughout the years. Between 1901 and 1905 a certain amount of inversion of the curve is noticeable, evidenced in the analysis by Ihe change of A a to the adjacent quadrant. From 1919 to 1920 a, slight amount of inversion may be perceived; the change of A., is in the opposite direction. Tin- sum of A.,'s in L904 and 1905 is 173°; in 1919 and 1920 it is 177°. Seasonal Vector Diagrams of Mean Diurnal Horizontal Disturbing Forces. These are published for 1910 and 1920 herewith, and require some comments. Those for 1905 and 1919 were published in the preceding year's report. Looking at these one is struck by the very great similarity in general shape of the summer diagrams in 1919 and L920, and the decided difference in the winter diagrams, from 21h. to 2h. (I. the winter diagrams do not differ much, except that the earlier hours are retarded on the diagram (or the variation accelerated), but from 16h. on to 20h. the difference is marked, and during tlese hours 1905 and 1910 winter diagrams correspond fairly .closely, while from I7h. to 20h. the general trend is somewhat the same, only somewhat smoother and inclined to the full rounded nature characterizing the whole winter diagram in 1919. in which year it indeed seems that the whole diurnal variation in winter was of simpler composition, and gives one the impression of a series of spirals, curves which I believe have in the past been unsuccessfully invoked in the interpretation of these phenomena. It is curious, however, that in many of these curves the position of the hours on the diagram gives one the impression of action in spirals. The slight constriction between lh. and 6h. evident in the 191!) diagram is also seen in the 1920 diagram. In the winter of 1910, 1919, 1920 the westerly disturbing force at about local midnight is large, and exceeds that at 23h. 6. to Oh. G., or local late morning. In the equinoctial night variation, 1910 much more, closely resembles 1919 than it does 1920. A striking difference between the winter vector diagrams in 1905 and 1919 occurs between 6h. and lOh. G. In both years this part of the curve is full and rounded, but tie convexity is directed inwards and south in 1905, and outwards and north in 1919. 2h. 0. seems to generally occupy the same position on the diagram, which means that at that time; of day the direction and intensity of the disturbing force has on the average for the season the most constancy in the different years. Visit of the " Carnegie." The ocean magnetic surveying-yacht " Carnegie," of the Department of Terrestrial Magnetism of the Carnegie Institution of Washington, revisited Lyttelt.on in October. The following account of her voyage from Fremantle is taken from the official report of Commander Ault, and after reading it one