variation at Scilly as being due to the variation of sea temperature,
but the sea temperatures are taken from means for a very short period,
these being the only ones immediately available, and it would require
a much more prolonged investigation of the sea data before a really
trustworthy diagram could be drawn. When these are obtained for a
number of stations it would not be difficult, by assuming the undis-
turbed climate to be the same for places in the same latitude, to
conduct a satisfactory analysis of the elements of temperature of any
place into its primary constituents. It is, however, apparent from this
mode of resolution of the temperature oscillation into components that
the more the components approach to coincidence of epoch the greater
is the resultant effect. For example, if the sea temperature round the
British Isles reached its maximum on June 21, the amplitude of tem-
perature at Kew would be raised to 13 0> 6 F. 5 and accordingly part of
the intensity of a Continental climate may be attributed to the simul-
taneous heating of the large area over which the air for a Continental
station passes.
There is another point of some interest which arises in connection with this subject, and that is, that all the successive stages of tem- perature change are delayed by the effect of the sea. Since both the first- and second-order components of the variation of sea temperature are of the same order of magnitude as those of the land curves, the effect of the sea is to delay the seasons, and we should on that account expect, not only the autumn and winter, but also the spring, to be later in a marine climate like that of Scilly or Falmouth than in an inland climate, and it would be interesting to ascertain how far such a conclusion is borne out by phenological observations. If this conclu- sion should be borne out, and there is at least some evidence in favour of it, one should seek an early spring, not on the coast, but inland, and on the other hand a late summer is more characteristic of the sea coast than of the inland country.
It may seem surprising that the effect of the sea upon the annual temperature ^scillation of a place near it should be regarded as increasing the amplitude of the oscillation, whereas it is a matter of common knowledge that the effect of sea upon climate is to reduce the amplitude of the temperature oscillation as compared with that of a Continental station. But in fact the temperature range of any place depends upon the original range (which may be regarded as depending merely upon its latitude) and the increase of its range due to its sur- roundings. In the case of an ocean station the increase of range is small, because the temperature range of the ocean is itself small, and the epoch of its maximum is two months later than that of the planetary effect. At a Continental station the increase due to the surroundings is much larger and more nearly simultaneous with the original planetary effect.
