and chromosphere are all projected into one point. Whether every man who has gone forth to solve the riddle of the corona has fully realized the odds against success is doubtful.
Much has been written concerning a possible eruptive origin, or about magnetic influences in shaping the forms of its streamers. It has been shown that the details of the corona at one eclipse are totally different from those at another, and that the outline form of the corona is a function of the sun spot cycle. At sun spot maximum the general form is nearly circular, and the polar streamers are nearly as bright as the equatorial streamers. At minimum, the polar streamers are much fainter than the equatorial ones, and long wings seem to extend out approximately from the spot zones. It is a surprising fact that, with all the changes of form, we do not yet know whether the materials composing the streamers are moving in, or out, or both, or neither. The epoch-making, large-scale coronal photographs by Schaeberle in 1893 opened a promising way of determining such facts, but astronomers have been slow in taking advantage of the opportunity. Photographs of the corona should be secured for this purpose at widely separated stations—preferably at three or more stations—with essentially identical instruments, and with equivalent exposures, in order that results may be as nearly comparable as possible. This effort to determine motion in the corona, it seems to me, is the most important problem of the coming eclipse; and, fortunately, the circumstances of widely separated stations in Labrador, Spain, Tunis and Egypt, and promising weather conditions at the last three are favorable for the attack. Considering all elements of the question, including that of probable unsteadiness of the atmosphere at one or more stations, the five-inch aperture, forty-foot focus cameras, promise the most directly comparable, and therefore the best, results. The only case of motion on coronal plates thus far observed seems to be that detected by Schaeberle, on the Chile-Brazil-Africa plates of 1893; and in this instance the moving mass was decided to be a comet, and not a part of the real solar appendage.
One of the most intensely interesting features ever observed in the corona was the tremendous funnel-shaped disturbance recorded on the Sumatra plates of 1901. Perrine was able to show, with essentially no room for doubt, that the vertex of the disturbance was immediately over the large and only sun spot visible on the sun in the week preceding and the week following the eclipse. The circumstances were unusually favorable for reaching this conclusion: there was but one sun spot; it was very near the limb at the time of the eclipse; there was but one region of unusual disturbance visible in the corona; this was on an extraordinarily large scale, and its vertex was near the sun's limb; and the disturbance and the sun spot had identically the same