A new whistler phenomenon has been identified through measurements at ground stations, on an Aerobee rocket between 100 and 200 km, and on the Alouette satellite at 1000 km. The new phenomenon is called the 'subprotonospheric' or 'SP' whistler, since most of its path appears to be restricted to the region below about 1000 km. The first example of an SP whistler was reported by Barrington and Belrose. In the present report a large number of observations are summarized, and the basic characteristics of the new phenomenon are described. Experimental results are presented which suggest that the whistler ray path is confined to the region between roughly 100-and 1000-km altitude, and that the whistler energy can echo back and forth between these levels. The SP phenomenon occurs mostly at night, typically within a few hours after sunset. SP events are often observed over a period of one or two hours in duration and, for a single Alouette pass, have been observed over a north-south range as great as 2000 km in extent. The evidence. suggests that the SP phenomenon occurs mostly near sunspot minimum and at dipole latitudes greater than 45 degrees.

In considering the propagation of lon g whistle rs and whistle r ec ho trains, the que stion ari ses about where the downcoming whistlers are refle cted. The se ve ral s uggestions that have bee n mad e include ground reflection and refl ection at the lowe r boundary of the ionosphere. In eithe r case, the ec ho of a daytim e whistle r would mak e se ve ra l more passes through the ab sorbing V region than the whi stle r itself, a nd we should expect whi stl e rs occurrin g a round noon to ha ve a much s maller probability of havin g ec hoes than whi stl e rs occurring at ni ght. An a nal ysis of se ve ral ye ars of data obt a ined a t the Da rtmouth Co ll ege whi stl e r stati on yield s the res ult , however, th at although th e a ve rage whi stl er rate is muc h hi ghe r a t ni ght than durin g th e da y, the proba bility of a whi stl e r having a n ec ho s hows littl e c ha nge fro m midni ght to midd ay. Co ns iste nt with thi s obse rva ti on a re th e res ults of ano th e r s tud y showing th at the diffe re nce in the inte nsity of a noo ntim e whi s tle r and its ec ho may be onl y a few decibe ls.If th e th eo reti cal predi cti ons a bout a bsorption of whi s tle r-mode waves a re even nearly correc t, our res ults on whi stl e r ec hoes a re in co mp a tible with the lo we r-bound a ry or ground·refl ec ti o n model. In no cases st udi ed by us has the whi s tle r echo bee n more inte nse th a n th e whi s tl e r itself, a nd we do not at prese nt favo r th e id ea th at whi stle r echoes a re a mplifi ed in th e magne tos ph er e. A mode l co nsis te nt with our results is one in whi c h a la rge fr ac tio n of the e ne rgy of a down co ming whi stl e r is refl ec ted above V·region heights . In thi s model a whi stl e r may be pi ctured as bounc in g bac k a nd forth betwee n th e ionosp heres of th e two opp os ite he mis pheres, with so me of th e e ne rgy " lea kin g through" to the ground a t one or both e nd s of th e pa th . Whi stle r obse rva tions could also be expla in ed by a model in whi c h t he daytime tra nsmiss ion loss fo r VLF e ne rgy is hi gh o nl y for th e firs t upwa rd pe ne trati on of th e ionosp he re, bu t s mall once th e e nergy is propagatin g in th e " whi stle r mode."