140 RECORD AND RESULTS OF 



Encyclopsedia Metropolitana, the variation of the mean level of the sea depends 

 upon the changes of the moon's and sun's declinations, but as the latter goes 

 through its changes in half a year, and as the zero levels of our two series are dis- 

 connected, we can only examine the lunar effect, which can be expressed by C sin "^S, 

 where the constant C amounts to a few inches to be determined by observation. 

 The constant C is greater in low and high latitudes, and very small in middle lati- 

 tudes. The oscillation wiU go through its changes in half a lunation (l4f days), 

 and we may expect high level at the greatest declination, independent of the sign, 

 and low level when the moon is in the equator. 



The breaks in our mean level readings, as examined above, sufficiently demon- 

 strate the insufficiency of the accuracy of our observations for so delicate an inquiry 

 as the variation in the mean level ; in some portions of the series the dependence 

 of this level upon the declination appears systematic, but is hidden in other por- 

 tions by irregularities. In Series I the mean of three readings of the level for ^ = 

 (after applying the corrections indicated) is 16.67, and for 5=+ 26° from two 

 readings is 16.88 feet, range 2| inches ; in series II the mean of three readings 

 (after applying the corrections indicated) is the same (17.80 feet) for 5 =0 and 

 ^ = + 25°, on the average therefore we would only have between one and two 

 inches of oscillation. 



But few investigations into the variations of the mean level have been made, 

 and more complete comparisons of observation and theory, on this point, are very 

 desirable. 



Effect of Changes in the AtmosplieriG Pressure upon the Tides. — Considering the 

 short series of observations any result for the dependence of the changes of the 

 height of the barometric column upon those of the sea level can only be a first 

 approximation, the result deduced from the observations is nevertheless entitled to 

 some confidence. The treatment adopted was the following : — 



The mean levels, each day, and for each series independently, were grouped in 

 two columns for days with barometer beloio, and for days with barometer above its 

 average value (30.01 inches for Series I, and 29.65 inches for Series II). The cor- 

 responding difference from the average value was also set down, and then the mean 

 of the whole series taken, thus : — 



For Series I, average level IG^'-Y, average depression of barometer 0'°.22 

 " " 1(>.6, " elevation " 0.24 



Or — 1 inch of change of level for 0'".46 of change of barometer. 

 For Series II, average level 18". 0, average depression of barometer 0'M5 

 " " 1Y.8, " elevation " O.lt 



Or — 2 inches of change of level for 0'".32 of change of barometer. 



From the two series combined we obtain therefore a change of — 3 inches for a 

 change of f inch (nearly) in the barometric column; in other words, a rise of one 

 inch of the barometric column will be accompanied by a corresponding fall in the 

 level of the water of four inches nearly. 



This result is also affected by any uncompensated part, by reason of the short 

 series of observations, of the effect of the variation in the mean level, and also of 

 the effect of the wind. 



