The position angle 2448 is 180° too great. for the measures can only refer to the star a. small.

It should be 64°-8,
The distance is too

A sketch accompanies the notes with the great reflector, p. 174. This shows the nebula and three stars α, β, y, following it. I have identified these stars with mine, as below :

a=a

β=b

y=c

I have computed, for comparison, the position angles and distances αβ, βγ and ay from my measures. They are :

I do not think these large residuals indicate any real change in the relative positions of the stars. In this case, the position angles with the large reflector seem to be fair, but the distances must be affected with very serious errors.

It will be seen that the presence and absence of the central star, in the Rosse observations, are strongly attested by the notes of 1857 Oct. 16, and 1873 Oct. 11.

Yerkes Observatory, Williams Bay, Wisconsin: 1908 February 6.

P.S., 1908 April 21.-In reference to the variability of the nucleus of the nebula N.G.C. 7662, I have carefully gone over all the literature of the subject, and in no case has there been any suggestion of variability (outside of my own suggestions). There does not seem to have been any suspicion, even, of variability.

[The half-tone is too harsh. This causes the central darkness to be a little too great, relative to the general disc. It also makes the bright, inner ring too bright. The south portion of this ring should be shaded or slightly darkened, as the north portion is somewhat the brightest part of the ring.-E. E. B.]

Note on the Period of Variation of Barnard's Variable Nebula in Andromeda. By Professor H. H. Turner.

At Professor Barnard's request, I add the following note on the period :

Selecting the well-determined maxima shown in the first column of the following table, the intervals can be represented by multiples of 27 days, as shown in the next three columns. In the last column is shown the approximate alteration of period when the multiple is altered by one whole unit.

It is clear that the long interval of 13781 days is of little use for defining the period at present; but Barnard's observations and the interval between Rosse and Lassell give about 27 days.

But the last column shows that, since all the differences are approximate sub-multiples of 2 days, periods of 29.3 or 25.3 days might suit the observations. Below are given the corresponding figures :

The accordance for the 29 day period is, however, not good.

The periods 274 and 25 days divide approximately into 355 days,

so that at the same time of year they run together nearly for several succeeding years.

Hence the observations do not readily discriminate between them; but those at 2416820, 2417176, and 2417178, when the nucleus was faint, seem to favour the 27 day period, since they make the epochs 10 or 12 days away from maximum, while the 25 day period makes them close to maximum. Professor Barnard proposes to make further observations to settle the question. :

Corrections to Professor Turner's Paper "On the Classification of Long-period Variable Stars, and a possible Physical Interpretation." By the Rev. J. Stein, S.J., Sc.D.

(Communicated by Professor H. H. Turner, D.Sc., F.R.S.)

[My best thanks are due to Father Stein for pointing out the corrections noted below. According to the general permission given in the last sentence of his letter, I publish it at once.-H. H. Т.]

Specola Vaticana, April 3, 1908.

Dear Professor Turner, - Lately I have studied your remarkable article on the "Long-period Variable Stars" (M.N., Ixvii. p. 332). My interest was the greater as this is the first serious attempt to bring the long-period light-curves in close relationship to the laws of sun-spot variation.

I take the opportunity for suggesting some ideas by which the number of possible cases in your hypothesis may be still more reduced.

1°. Is there not a little oversight in your paper on page 345 There it is said"The mean (foreshortening) factor for the spot may be taken as

Then "the total effectiveness" would be (top line of page 346)—

I s = (4ẞ cos e sin x + 4 cos ẞ sin e cos x),

2π

and in the "polar view".

81 = sinx (as correctly stated on page 342, line 6; but not as incorrectly on page 346).