THIS essay contains a full discussion of all the available observations of a remarkable variable star, the fluctuation of whose light presents many features the explanation of which is still beset with much difficulty. Of the 12,000 recorded comparisons of the relative brightness of η Aquilæ and neighbouring stars, no less than 7147 are due to the indefatigable perseverance of Julius Schmidt, who for the last twenty years of his life (1859-1879) was director of the observatory at Athens. Taking advantage of the preponderance of Schmidt's records over those of all other observers, the author first, investigated the variations of the star's light as shown by these alone, and then examined in how far the results thus obtained were confirmed by the labours of other observers. The discussion of Schmidt's observations was complicated by the circumstance that in the majority of cases he only used two comparison stars, β and ι Aquilæ, one of which there was great reason to believe was itself variable. From the 1700 occasions on which both stars were used, Dr. Lockyer found ι to vary 1˙7 grades of Schmidt's scale with a period of about thirty-five years. Associated with this long-period change, the relative brightness of ι was shown to be subject to an annual variability of about one grade, which is clearly traced to the influence of the hour angle at which the stars were compared. This apparent annual variation is, therefore, identical with the “position error” noted by many observers of variable stars, and which has lately been investigated so thoroughly by Mr. Roberts, of Lovedale, South Africa. Schmidt's observations, freed from both the above variations as well as from the slight aberration error due to the changing position of the earth iaits orbit, were all reduced to the meridian of Bonn. Owing to the general sparseness of the data, it was necessary to combine the observations of 100 periods, or about two years, into a group represented by a single curve. This was effected with the help of a provisional mean period and date of minimum for which Argelander's values of 7d. 4˙23444h. and 1848, mag. 18d. 6˙4333h. M.T. at Bonn were adopted, the date being that of epoch 400. To further facilitate the construction of the curve the mean brightness in each 6-hour interval of the period was set down for each group, together with the corresponding time and the number of observations. In this way the whole of the materials collected by Schmidt from 1844 to 1879, and tnus covering eighteen groups of 100 periods each, are expressed in tabular form on p. 21 of Dr. Lockyer's dissertation. The numbers for each period were then plotted on millimetre paper; tentative and afterwards definite curves representing the fluctuations of the star's light were drawn through the points thus obtained. The resulting eighteen curves representing Schmidt's observations are reproduced one above the other on a small but distinct scale on Plate I., each curve being prolonged by about one-third of the period to bring out the form of the minimum. A study of these curves shows that Argelander's period cannot, at present, be improved on as a mean value. The time of Argelander's epoch 400 is, however, shown to be too early by oan interval which the author fixes at 3˙2655 hours. The curves also show conclusively that the epoch of maximum oscillates to and fro to the extent of five hours on either side with a period comprising 400 maxima. In the same way the minima are subject to an oscillation of about three hours in approximately 2400 periods. The combined effect of these variations causes the light-period to “vary between 7d. 4h. 14m. 40s. and 7d. 4h. 13m. 28s. also in 2400 periods. Superposed on the main light curve are four subordinate undulations, of which more hereafter.