A few words on the subject of double lights, naturally spring out of what has been said about the masking of lights. The term _double lights_ is properly and distinctly confined to lights on different levels, but not _necessarily_ (as _leading-lights_ are) in separate towers. The sole object of using double lights is for _distinction_ from neighbouring lights; and they are unquestionably most effective in this respect, when they are placed in the same tower. In this point of view, therefore, I shall speak of them; and it is obvious that all that peculiarly belongs to them is, that the difference of level between them shall be sufficiently great to present the lights as separate objects to the eye of the seaman, when placed at the greatest distance at which it may be desirable that he should be able to recognise their characteristic appearance. In many cases it is not necessary (but it is certainly always desirable) that the lights should, from the first moment of their being seen, be known as _double lights_; but in others, it may well consist with safety, that two lights, which appear as a single light when first seen at the distance of 20 miles, shall at 15 or 10 miles distance be discovered to be _double_. Now we should at first be apt hastily to imagine, that all that is required to produce that effect, is to raise the one light above the other to such an extent, that the distance between them shall be somewhat more than a _minimum visibile_ at the most distant point of observation, or, in other words, that the difference of the height of the lights should be such as to subtend to the eye at the point of observation, an angle greater than 13″·02, which is the subtense of a _minimum visibile_ during the day.[80] But the effect of irradiation, to which I have already alluded, tends to blend together the images of the lights long before their distance apart has become so low a fraction of the observer’s distance from the Lighthouse, as to subtend so small an angle; and I have accordingly found by experiments, conducted under various circumstances, and at various distances, that repeated observations gave me 3′ 18″ as the mean of the subtenses calculated in reference to the distances at which the lights began to be blended into _one_. [80] This quantity is deduced from observations made by my friend Mr JAMES GARDNER, while engaged on the Ordnance Survey, and may be regarded as the _extreme_ limit of visibility, under the most favourable circumstances as to the state of the atmosphere and also the contrast of colours. The observed object, also, was a pole, not a _round_ disc; and it is familiar to every one accustomed to view distant objects, that _vertical length_ is an important constituent in their visibility.