can be seen most easily and most distinctly in the place where the bows are, but they absolutely cannot be seen anywhere else.
The parallelism also of the streaks to the bows, in the author’s estimation, proves that the same cause which determines the direction of the how, must determine that of the streaks, and thus establishes their dependence on critical separation.
Dr. Herschel also contends, that streaks of different colours could not be produced by a plate of air, of uniform thinness, between plain surfaces, and that the prevalence of a blue colour in the streaks be- longing to the blue bow, and of the converse in those belonging to the red bow, prove their dependence on critical separation.
Since it has been conceived by other persons, that by means of a plate of air, havingtheform of an extremely thinwedge, straight bands of colour would be produced between plain surfaces slightly inclined to each other; and as an experiment in support of this opinion had been shown to the author, he gives his own explanation of the fact : and he ascribes the production of the colours to distortion of the surfaces, because a degree of force was in that instance employed for the purpose of producing the requisite contact at one extremity of the glass. And since in other experiments, made with perfectly plain surfaces, where no pressure was employed no streaks could be seen, Dr. Herschel concludes that when streaks are seen, the surfaces employed are either not plain in their general extent, or are terminated by some inconceivany small curvature at the edges in contact.
It has, in the next place, been observed to the author, that in the enlarged figure which he has given in his last paper to illustrate the streaks, the vacancies observable correspond with, and depend upon, the assumed intervals between the rays, which in that figure are re- presented as originally separated by blank spaces.
Dr. Herschel admits that there is some plausibility in this objection, but informs us, that the supposed force of it is founded on a misconception of the figure, which is not designed to represent the visible arrangement and colours of the streaks, which can only be deduced from their mixture at the place where they enter the eye ; but he declines a thorough investigation of this point, because it would really be an endless undertaking.
One section of the present communication is devoted to the consideration of the breadth of the streaks compared to that of the bows, and the cause why they must take up a broader space than the bows from which they are derived; because it has been remarked, that this circumstance precludes the possibility of accounting for them by critical separation. But although this remark may at first view appear to be justified, it must be remembered (says the author) that the modifying power of the surfaces is added to the principle of the critical separation. The modification specifically named by the author, is that of reflection by the plain surface held under the prism, which, in the first instance, magnified the extent to 2% times the breadth of the bow ; and if the reflection be repeated any number of times be-
