226 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1956 



lines of magnetic force will squeeze the mass and eject it into space. I 

 have called this type of action the "watermelon-seed effect," because 

 the ejected lump is something like a slippery watermelon seed held 

 between the fingers and squeezed. Note especially that the imparted 

 velocity may be many times greater than that of the moving fingers. 

 In the case of the sun, the resultant jet may very well be supersonic. 



The magnetic fields of the sun must exert a profound effect on the 

 flow of gas through the corona and prominences. In regions of low 

 density and strong field, the material will generally be constrained to 

 move along the lines of force. But in regions where the density is 

 higher, the ionized gas will sweep the lines of force along with it. 

 Thus the contention made by some astronomers that the flow patterns 

 of prominences provide a map of the solar magnetic field is not gener- 

 ally true. Often, as for the loop prominences described above, the 

 flow is along the lines of current rather than of magnetic force. 



From time to time we observe, usually near sunspots, areas that 

 shine particularly brilliantly in the light of hydrogen. The physical 

 nature of such eruptions has been the subject of discussion for a long 

 time. Observation shows that the flares usually start as one or more 

 brilliant points of light, growing rapidly to delineate a veined net- 

 work, which is apparently merely an enhancement of the characteristic 

 reticular structure that existed before the outburst. Although con- 

 vection may have been responsible for the initial ejection of heated 

 gases, the growth from bright points into linear structures strongly 

 suggests that flares seen against the disk are loop prominences seen 

 at the limb. In fact, the most famous of all flares, which occurred on 

 September 1, 1859, clearly suggests such an association. 



Two independent observers, Carrington and Hodgson, simultane- 

 ously saw two brilliant points of light appear within an active sunspot 

 group they were sketching. According to the former observer, the 

 spots rapidly developed into the shape of a crescent. These observa- 

 tions were in white light, not in Ha. Hence the flares must have been 

 extremely intense. Their surface brightness, according to estimates, 

 was at least five or six times that of the neighboring photosphere. 

 Within five minutes the outburst had disappeared. There is no doubt 

 in my mind that the "crescents" reported were loop prominences seen 

 from above. However, even though the flare itself represents a rela- 

 tively stationary phenomenon in the solar atmosphere, extending its 

 growth by propagation along definite paths, observation indicates 

 that some material may be ejected. Dark absorbing clouds, often 

 moving at high speeds, have been seen to accompany flares. It is too 

 early to say, however, whether this represents an effect of the flare as 

 such or is merely an outlying portion of the ejected matter that was 

 later condensed to form the flare. In certain instances we have iden- 



