If we compare the mean errors with the observed alterations in longitude, we see that the latter are considerably greater. On that account Koch concluded[1] that: “It appears from the preceding that the sources of error given are neither singly nor collectively sufficient to suggest an explanation of the difference of about 1190 m. which exists between the position of Haystack, as determined by the Danmark-Ekspedition and the Germania-expedition. The only source of error, which in this connection plays any part, is the astronomical determination of longitude. But in order to explain the deviation through the error of the longitude of the observatory, we must put the actual error of the astronomical determination of longitude at four to five times the mean error, which is perfectly absurd.” F. Burmeister,[2] it is true, has objected to this on the grounds that the mean error only plays the part claimed for it, if there are an infinite number of observations, and that in this case the actual error of the result might reach the amount of the observed difference. Therefore he does not consider Koch’s proof as sufficient. Even if this is correct from the theoretical standpoint, and we cannot rely on the result obtained, but must attempt to obtain new and more accurate measurements by the aid of wireless telegraphy, I still believe that Burmeister’s criticism overshoots its mark. If the exact quantitative proof must be reserved for more accurate measurements,
