258 
IONISATION, DIFFUSION, ROTATION 
Table 34 contains Fowler and Guggenheim’s results for a typical star 
at various distances from the centre. The star chosen is of mass 2T3 and 
mean density -002, that is to say, a rather faint giant star of about type G. 
The first column gives the distance from the centre in terms of Emden’s 
unit B/ 6-9. 
Table 34. 
Variation of Molecular Weight in a Star. 
z 
T x 10- 6 
P 
Molecular weight for 
Oxygen 
Iron 
Silver 
0 
6-59 
•1085 
1-95 
2-33 
2-87 
i 
5-64 
•0678 
— 
2-33 
— 
2 
3-84 
•0215 
1-93 
2-34 
2-94 
3 
2-37 
•0050 
1-92 
2-53 
3-37 
4 
1-38 
•0010 
1-95 
316 
4-0 
Main series 
26-4 
6-95 
1-97 
2-18 
2-40 
It will be noticed that /x is rather higher than appeared from the crude 
calculation. Iron even at the centre of the star retains 3 electrons, whereas 
we have previously allowed it 1| electrons; this is a rather large difference 
since there is a big gap in ionisation potential between the second and 
third electrons. Oxygen retains § of an electron against nil on the crude 
theory. These changes are mainly due to the corrections (a) to (e) and not 
to the difference in the temperature and density. 
Judging from these figures the formula /x oc T~' r used in § 13 (see also 
§ 94) will, if anything, overcorrect the variation of molecular weight 
within a star. 
In the last line of the table we give the mean molecular weight (i.e. at 
a place where T is § the central temperature) for a star of the same mass 
which has reached the main series. It happens that both Ag and Fe are 
critical elements for this change, and most elements would have shown a 
much smaller difference. There appears to be no reason to amend our 
former conclusion that the range of mean molecular weight between Capella 
and Sirius-Sun conditions is not likely to exceed 0-1. 
Table 35 extends the comparison to stars of different mass on the main 
series. Of particular interest are the results for stars of small mass. We 
felt some hesitation in applying the crude theory to these owing to the 
lack of cleanness in the ionisation. 
The increasing overlapping of the ionisation is exhibited in the last 
columns of the table, which shows the percentages of iron atoms retaining 
respectively 1, 2, 3, 4 electrons. The change of ¡jl is remarkably small; 
but that is partly because the elements chosen are not in a critical state