530 



NATURE 



[April lo, 1890 



Some grounds must be given for the severe stricture 

 here passed, and the only difficulty is to select the most 

 fitting examples from so much worthless matter. On 

 p. 18 the author says : " It is considered sufficiently near 

 to the truth, if our calculations came within a fc^v hours 

 of the time and near enough to the quantity of the eclipse 

 to identify it as being in all probability the obscuration 

 mentioned by the historian in connection with a certain 

 event." The italics are our own, and the statement to which 

 they call attention is absolutely a misrepresentation. It 

 is scarcely necessary to say in these columns that no 

 astronomer of repute would be satisfied with a dis- 

 crepancy of anything like this amount between history 

 and computation in any case in which the phenomenon 

 is clearly indicated and accurately described. In the 

 annexed table is given the comparison of the computa- 

 tions of various astronomers of the times of historic 

 eclipses with the recorded times. To keep the table to 

 a moderate length it is confined to those dates between 

 which the examples have been worked out by the writer. 

 In estimating the accuracy of representation, there are 

 two circumstances to be taken into account. One is 

 that an eclipse, being a phenomenon the exact time of 

 whose occurrence could not be accurately predicted by 

 the observer or recorder, must have been in progress 

 some time before detection, or, all observations of the first 

 geometrical contact, the phase computed from the tables, 

 would be observed too late ; and though the error from this 

 cause would not be so large in the observation of the end 

 of the total phase, it is probable that this phenomenon 

 would be recorded too soon. The other circumstance is 

 that we cannot regard Ptolemy, from whose work the 

 times here given have been taken, as a totally unpre- 

 judiced witness. He was anxious to establish a theory, 

 and it is probable that he selected those instances which 

 most nearly fitted his preconceived system. In other 

 words he may have — what is not unknown in these days 

 — rejected a discordant observation. 



It is needless to point out there are no discrepancies 

 of a few hours between the tabular and observed facts, 

 and that the grave charge of the lack of accuracy is un- 

 sustained. The circumstances of two of these eclipses 

 have been worked out by the author with some pretence 



of detail, employing his '•' new and corrected tables." For 

 these two echpses, - 382, Dec, and - 200, Sept., he gives 

 the London mean times of the true full moon i5h. 56m. 

 and 3h. i6m. respectively. There is no attempt to deter- 

 mine the e.xact phase observed, and it may be remarked 

 that the longitude given for Babylon is grievously in 

 error. Tliese two eclipses have been selected with the 

 particular purpose of demonstrating that no secular 

 acceleration of the moon's motion exists. This selection, 

 with this view, is unhappy. With regard to the earlier 

 eclipse, it is very doubtful if it was really seen at 

 Babylon. The account given in the " Almagest " 

 ("Halma," p. 275) rather suggests that Athens, or one 

 of the Ionic colonies, was the place of observation, since 

 the description of the date is by means of the Greek 

 calendar ; and Hipparchus says that this eclipse with the 

 two immediately following are added to the catalogue of 

 the Babylonian eclipses as though they had been observed 

 in that place (ws tfcct TfTrjprjfievas yeyovivai). This sugges- 

 tion that the record of the eclipse was made elsewhere 

 than at Babylon is strengthened by the addition of the 

 note that "the moon set eclipsed." In an eclipse which 

 commenced only half an hour before the setting of the 

 moon, these words would have little meaning, but if the 

 note was added by the observer at Athens, its purpose is 

 intelligible, for the eclipse would be more than half over 

 before the moon touched the horizon. It is very possible, 

 therefore, that some allowance for longitude was made by 

 Hipparchus, but with such a doubt overhanging the re- 

 corded time of observation, the selection of this eclipse 

 from the long catalogue collected by Ptolemy gives a very 

 doubtful support to any hypothesis. The second eclipse 

 quoted was doubtless observed at Alexandria, but if 

 Hipparchus is correctly rendered by Ptolemy, he is made 

 to say that the eclipse began half an hour before the 

 moon rose. The record, therefore, refers to a calculated, 

 and not an observed, phenomenon, and on that ground 

 alone should not have been selected. 



But it is in solar eclipses, the total phase being confined 

 to a comparatively narrow zone of country, that the 

 feebleness of the author's method is most conspicuously 

 exhibited. The eclipse known as that of Xerxes will 

 serve for an example. To adequately explain the cir- 

 cumstances as recorded by Herodotus and Aristides has 

 exercised the ingenuity, but baffled the efforts, of many 

 experts. It offers no difficulties to Mr. Page, though we 

 cannot think that his rendering will be generally appre- 

 ciated. Herodotus's description runs, " The army having 

 come out of their winter-quarters in the opening of spring."' 

 In the latitude of Sardis the opening of spring could 

 hardly be put as late as April 18, but this is the date 

 selected by Mr. Page, because on that day -480 there 

 was undoubtedly a total eclipse of the sun. The writer 

 does not mention, what is equally the fact, that the shadow 

 of the moon first touched the earth in the Indian Ocean, 

 passed over the Himalayan peninsula, through China, 

 and disappeared in the Pacific. Such a path is totally 

 inadequate to explain the further description of Herodotus, 

 that " night came on instead of day." 



A still greater absurdity is introduced when the autl 

 wishes to prove that the death of Augustus happened! 

 the year 13, by means of a solar eclipse which is saidj 

 have occurred iust before the death of that Emperor. 



