12. Distinctions of lights, founded upon the minute estimation of intervals of time between flashes, and especially on the measurement of the duration of light and dark periods, are less satisfactory to the great majority of coasting seamen, and are more liable to derangement by atmospheric changes, than those distinctions which are founded on what may more properly be called the _characteristic appearance_ of the lights, in which the times for the recurrence of certain appearances differ so widely from each other as not to require for their detection any very minute observation in a stormy night. Thus, for example, flashing lights of five seconds interval, and revolving lights of half a minute, one minute, and two minutes, are much more characteristic than those which are distinguished from each other by intervals varying according to a slower series of 5″, 10″, 20″, 40″, &c. 13. Harbour and local lights, which have a circumscribed range, should generally be fixed instead of revolving; and may often, for the same reason, be safely distinguished by coloured media. In many cases also, where the purpose of guiding into a narrow channel is to be gained, the leading lights which are used, should, at the same time, be so arranged as to serve for a distinction from any neighbouring lights. 14. Floating lights, which are very expensive and more or less uncertain from their liability to drift from their moorings, as well as defective in power, should never be employed to indicate a turning point in a navigation in any situation where the conjunction of lights on the shore can be applied at any reasonable expense. ~Height of Lighthouse Tower, and its relation to range of Light.~ The spheroïdal form of the Earth requires that the height of a Lighthouse Tower should increase proportionally to the difference between the Earth’s radius and the secant of the angle intercepted between the normal to the spheroïd at the Lighthouse and the normal at the point of the light’s occultation from the view of a distant observer. The effect of atmospheric refraction, however, is too considerable to be neglected in estimating the _range_ of a light, or in computing the height of a Tower which is required to give to any light a given range; and we must, therefore, in accordance with the influence of this element, on the one hand _increase_ the range due to any given height, and _vice versa reduce_ the height required for any given range, which a simple consideration of the form of the globe would assign. In considering this height, we may proceed as follows:-- [Illustration: Fig. 92.] Referring to the accompanying figure (No. 92), in which S′ _d_ L′ is a segment of the ocean’s surface, O the centre of the earth, L′L a Lighthouse, and S the position of the mariner’s eye, we obtain the value of LL′ = H′, the height of the tower in feet by the formula, 2 _l_² H′ = ------ (1.) 3