The "spreading" phenomenon above isn't the "black drop effect". It may be an ether drift effect, caused by a thin layer of ether flowing around the moon. If so, the amount of spreading of a planet at occultation might depend on the lunar phase, because the lunar phase, determines the ether velocity due to Earth's orbital motion.
I found three published photos showing displacement of a lunar peak seen at the moon's edge, but to the accuracy of my measurements (with a mm ruler, sometimes by eye, from a book page) these seem to be explained by libration. I think the peak that I studied, is on the E (farside) or maybe SE rim of the Lyot crater. See Z Kopal, "A New Photographic Atlas of the Moon" (Taplinger; New York, 1971) Plate 148 p. 219; and GL Gutschewski et al, "Atlas & Gazetteer of the Near Side of the Moon" (NASA; 1971) SP-241, Plate 5-3 p. 14, Plate 38-1 p. 22. The peak is at about 50S 88E or maybe 53S 87E. (South of the main peak, is a longer, lower, double, peak, also visible at the moon's edge in Alter's lunar atlas, Plates 40 & 41, but past the edge of the frame in de Callatay. Also someone had sliced out de Callatay's photo of a Soviet moon probe's landing site.)
I drew a line through the northernmost edges of the interior shadows of two reference craters, Fraunhofer V & B. Fraunhofer V prominently straddes the nearside rim of the crater Fraunhofer. Fraunhofer B is a prominent, isolated, simple, round crater about 2/3 the diam. of Fraunhofer; its N shadow edge is at about 42S 67E. See Gutschewski, Plate 52-2 p. 35. See also Plate 523, "Lunar Orbiter Photographic Atlas of the Moon", DE Bowker & JK Hughes, NASA SP-206, 1971. A drawn USAF Mercator moon map is reprinted in FM Branley et al's "Astronomy" (1975); and in Kopal, op. cit.
D Alter, ed., "Lunar Atlas" (Dover; New York, 1964) Plate 41, shows the moon at phase 4.59d in 1938. V de Callatay, "Atlas of the Moon" (Macmillan; 1964), Plate 1, p. 98, shows the moon at phase 3.53d on Dec. 30, 1943. Alter gives the libration angles and time of observation; these agreed with the American Ephemeris. For de Callatay, I got the libration angles by assuming a 6PM observation time at Pic du Midi (roughly consistent with the terminator line, i.e. solar colongitude, compared to Alter Plate 40, and with good sky position) & interpolating in the American Ephemeris.
I considered the libration as the sum of infinitesimal rotations called "to/from" (the peak moves toward or away from Earth) and "twirling" (the moon rotates about the axis defined by the peak). Because the peak will be taken to be at 45S latitude (a good approximation for "twirling" because, luckily, the long. & lat. librations are about equal for all three Plates considered), these are easy to calculate from the regular libration angles.
"To/from" libration rapidly changes the identity of the apparent peak, by bringing other parts of the ridge into position, but luckily this libration, whether the peak is taken at 45S or 50S, differs only about a degree between the plates considered. The smallness of the differences in "to/from" libration is confirmed by measurement on the photos, using the NS length of crater Furnerius vs. its distance from the moon's edge. The identity of the peak moves only reluctantly with "twirling" libration. Because the reference craters make a line almost coplanar with the twirling axis and the Earth, to/from libration, to first approximation, doesn't disrupt their collinearity with the peak.
It appears from the 1938 American Ephemeris, that the libration published in Alter, Appendix A, is for Washington D.C. For an observation at 8PM Pacific Standard Time, of a 4.6-day-old moon, the 3000 mile distance across the US would project to half that, i.e., roughly 3/8 degree greater libration in longitude. For de Callatay's presumed 6PM observation of a 3.5-day-old moon, the 5000 mile transatlantic distance would project 1/sqrt(2), i.e., roughly 7/8 degree less libration in longitude.
So we have librations for Lick Plate #2 ( = Alter Plate 41) of long. +5.54, lat. +6.63; and for de Callatay long. +1.21, lat. +2.31. The difference in "twirl" libration is (5.54+6.63-1.21-2.31)/sqrt(2)= 6.12 degrees. (Direct measurement on the photos yielded a corroborating estimate of 5.8 degrees.) Spherical trigonometry approximations gave 33km expected displacement of the peak (northward in Alter Plate 41, vs. deCallatay), along the moon's edge, referred to the line through the reference craters where it intersects the moon's edge. The actual displacement is measured to be 31 km, using 57 km plateau-to-plateau NS diam. of the crater Fraunhofer as a yardstick.
Alter's Plate 40 (Lick Plate #1) requires about +3/4 / sqrt(2) degree geographic correction in long. lib., giving long. lib. +3.93, lat. lib. +3.76. The same peaks appear as on Alter's Plate 41. (3.93+3.76-1.21-2.31)/sqrt(2)= 2.95 degrees, giving 16km expected displacement northward vs. deCallatay, relative to the reference craters' line; measured is 13km.
Topic Author