CompositionPri01" Lo 1961 it was gener ally accepted tha t the main cons Lituen t of the atmosph er e above 300 km was atomic oxygen up to a n altitude of about 1000 km and that at hi gh er altitudes hydroge n would predominate. Direct ex perim ental evid e nce for th e presen ce of neutral h ydroge n in the outer atmosph er e has come from t he high resolution Ly m an a spectrum obtained by Purcell a nd Tousey [1960]. From th e absorption core of th eir L ym an a spectrum , the total con ten t of hydrogen above the altitude of the rock et measurem ent has b ee n determin ed . John son [1961] has interpreted t his hyd roge n co ntent to be cLi stributed in t he form of n, "geocorona." The problenl of interpretation of the L y man a observations in terms of hydroge n and its location wi th r espect to the earth has r ecently b een r evi ewed in detail b y Donahue [1962]. The computation of the distribution of neutral h ydrogen is complicated by the fact that at magn etospheric levels "exospheri c" conditions prevail, i. e., the mean free path of neutral hydrogen is greater than the local scale height. Because of the escape of particles with high velocities, the velocity distribution is, strictly speaking, non-Maxwellian and the distribution of density with hei gh t does not follow t he simple h ydrostatic equation, but has to be computed by taking into account ballistic escape, b allistic r eentry a nd bound-orbiting particles. Such compu t a.tions have been mad e by bpik and Singer [1961] and by Johnson [1961]. Out to a geocentric distance of a bout two ear th r adii , a simple h y drostatic dis tri bution , however , stillreprese n ts fI, r easonably good n.p pro xi III ation .Photoionization and /or ch ar ge exchange with oxygen ion s of the neu tral hyd rogen of th e geocoron a lead to the protons constituting t h e " protonosph er e." rrhe ions fLt J1h1gn etospheric levels are distribu ted according t? a diffu siv e equilibrium distribution , i.e., hydrostatIcally supported bu t constrain ed by the ~arth 's magn etic field. For ions the m ean fr ee pn.th l S short enough so that the concep t of an ionexosphere is not applicable [Johnson , 1962] . While the diffusive equilibriurn distribution of an ionic sp ecies ,~as gen~]"ally con~id ered to be go:v:erned by a scale h eJght t WICe t hat of t h e cOl"l"espondmg neutral sp ecies, it was pointed out by Mange [1960Mange [ 1961] that t his con cept is not justified for a minor ' ion in tbe presen ce of other ions. The electric field which is set up to prevent further charge separation ' of electrons and ions diffu sing under gravity, and which is propor tional to t h e me}1n mass of the (s ingly charged) positive ions, ca uses the density of minor ligh t ions first to increase with alt it ud e until it b eco mes predominant, n.Iter whi ch it shows the llsual expon ential decrease with altitude according to n. scn.le h eigh t appr~aching twice that of t he conesponding n eutral constItu ent. S ubsequently to . Mange's independent derivation of the equation go...