There is no such thing as a unitary, or uniform, "green flash" phenomenon, such as you may have read about in less authoritative sources. Let us stick to what we know. The atmosphere separates photospheric light like a prism does. Thus, the violet sun rises first and the red sun last. At sunset the opposite is true, the red sun setting first and the violet sun last. Thus, in its first moments at sunrise and its last moments at sunset, the color of the Sun will tend to vary (from its color when all of the "colored suns" are seen simultaneously, as is the case nearly all day). Superimposed on these color changes is the color change from the long path taken simultaneously through the atmosphere. This long path tends to scatter blue light more efficiently (thus the blue color of the midday sky) making the image of the Sun, which is formed only from light which survives the trip, redder in color. In particular, violet and blue are frequently seen to be missing altogether. The long path also quite commonly facilitates a wide variety of meteorological modifications of the image. My observations reveal a wide range of phenomena which seem well-explained by these basic principles. In particular, the green flash is not something you either see or don't see, like some sort of step-function. I have seen the classically described phenomenon, an extremely bright flash of green light, as "bright as flashbulbs". Much more frequently, I have seen a bright green glow, lasting a greater or lesser length of time, and assuming a variety of shapes and sizes. More frequently than that, I have seen merely bright green specks, also of various durations and sizes. The large number of these specks I have seen may be inflated by the preponderance of my observations over a land horizon at sunrise through binoculars, where it is easy to catch a "first bit" through leaves of trees. Many, many times, there is no green, not really, just yellow, or orange or red; many of these, of every hue. I generally ignore any dim phenomena at sunrise, since I can't distinguish them from isolated illuminated clouds. And it is only with experience that I can properly anticipate the pace of events at sunset and make this distinction then. On extreme occasions, the phenomenon is not green, but blue. I have seen a number of bluish-green and greenish-blue phenomena, of various hues, mostly very green. At the equinoctial sunrise, 1998 SEP 23, I saw a BRIGHT BLUE FLASH. In detail: The Sun rose behind trees, which are far away and not of any great consequence. Immediately prior to sunrise, the whole area of the sky near the orient was especially luminous, so much so that it raised the question "Why is the sky already so bright if the Sun isn't up yet?". I saw a definite brightening of the pale orange at the orient. 31.40 seconds after that, at 11h 18m 23.6s UTC (I was too excited to have the patience to obtain the hundredths, personal equation = about .25s not applied), I saw the "first bit"; it was bright green, rapidly evolving, giving way to green and yellow spreading out horizontally, and then it happened. A BRIGHT BLUE FLASH, as blue as could be, not more green than violet, occurred 6.26s after the "first bit" and high, not directly in front of the first bit; but a fraction of a degree higher, and seemingly big, perhaps as big as the, as yet unrevealed, Sun. After the blue flash, the color continued to evolve, moving left and right, until the green ended 3.24s after the blue flash. I spoke with an experienced and enthusiastic meteorologist, Andre Bernier, who said there was nothing very unusual about the weather at sunrise. He said there was a small inversion of 10 or 15 degrees F for about 200 feet or 1000 feet at the surface. Apparently, the key to seeing blue phenomena is a clear path, which can occur even if there are a few clouds as there were. As with my 9 observations of the corona of our Sun from within the umbral shadow of the Moon, but to a lesser degree, the words above are rather definitely insufficient to thoroughly describe the phenomena. There is something very significant missing. You have to be there and see them yourself to really understand. (And this remains certainly true even if right now you are smugly telling yourself I am surely mistaken in this. It's an observing thing). Cheers. Walter Nissen dk058@cleveland.freenet.edu -81.8637, 41.3735, 256m elevation --- This sky observing report is pretty far off-topic. But I know from direct correspondence with a number of you that you would like to hear this from me. So long as no one without such special knowledge emulates this excursion, we should be ok. Please note, especially, that the purpose of this message is to share desired information, and not, e.g., for some selfish purpose. Even saying that, I apologize to those more focused on satellites.