26 THE VOYAGE OF H.M.S. CHALLENGER. 



and choosing any convenient hour for commencement, let the height of the water he ohserved at the following 

 times, reckoned from the commencement : — 



The ohservations may he regarded as forming three groups of three each, the memhers of each group being 

 separated by 8 hours solar or lunar, while one group is separated from the next by 8 hours lunar or solar. In 

 the mean of the nine results the lunar and solar semi-diurnal and diurnal inequalities are all four eliminated. 



Nine is the smallest number of observations which can form a complete series. If the solar day be divided 

 into m and the lunar into n equal parts, where m and n must both be greater than 2, there will be mn observa- 

 tions in the series ; and if either m or n be a multiple of 3, or of a larger number, the whole series may be 

 divided into two or more series having no observation in common, and each complete in itself. The accuracy 

 of the method can thus be tested, by comparing the means obtained from the separate sub-series of which the 

 whole is made up. 



Should the ship's stay not permit of the employment of the above method a very fair determination may 

 be made in less than a day, by taking the mean of n observations taken at intervals of the wth part of a lunar 

 day, n being greater than 2. Thus if n = 3, these observations require a total interval of time amounting to 

 only 16 h 32 m . The theoretical error of this method is very small, and the result thus obtained is decidedly to 

 be preferred to the mere mean of the heights at high and low water. 



The mean level thus determined is subject to meteorological influences, and it would be desirable, should 

 there be an opportunity, to redetermine it at the same place at a different time of year. Should a regular series 

 of observations for a fortnight be instituted, it would be superfluous to make an independent determination of 

 the mean sea level by either of the above methods at the same time. 



Besides taking observations on the ordinary waves of the sea when at all remarkable, the Scientific Staff 

 should carefully note the circumstances of any waves attributable to earthquakes. 



Specific Gravity. — The specific gravity of the surface and bottom water should be carefully compared, 

 whenever soundings are taken ; and whenever serial soundings are taken, the specific gravity at intermediate 

 depths should be ascertained. Every determination of specific gravity should be made with careful attention 

 to temperature ; and the requisite correction should be applied from the best Table for its reduction to the 

 uniform standard of 60°. It would be well to check the most important results by the balance ; samples being 

 preserved for examination in harbour. Wherever the temperature of the surface is high, — especially, of course, 

 in the Intertropical region, — samples should be collected at every 10 fathoms, for the purpose of ascertaining 

 whether any effect is produced upon the specific gravity of the upper stratum by evaporation, and how far down 

 this effect extends. 



Transparency of the Water. — Observations for transparency should bo taken at various depths and under 

 different conditions by means of Mr. Siemens' photographic apparatus. As, however, the action of this' depends 

 upon the more refrangible rays, and the absorption of these and of the more luminous rays might be different, 

 and that in a manner varying with circumstances, such as the presence or absence of suspended matter, &c; the 

 transparency of the sea should also be tested by lowering a white plate or large white tile to various measured 

 depths, and noting the change of intensity and colour as it descends, and the depth at which it ceases to be 

 visible. The state of the sky at the time should be mentioned, and the altitude of the sun, if shining, roughly 

 measured, or if not shining, deduced from the time of day. 



Relation of Barometric Pressure to Latitude. — In Poggendorffs Annalen, vol. xxvi., 1832, p. 395, is a 

 remarkable paper by Professor G. F. Schouw on the relation between the height of the barometer at the level 



