Juny 12, 1918] 
agencies of wind, ocean currents and birds. 
About 80 per cent. of the native land plants 
inhabit the West Indies, or southern Florida, 
or both. About 8.7 per cent. of the total 
native flora is endemic, there being 61 species 
in Bermuda, or its waters, not known to grow 
naturally anywhere else in the world. These 
plants are of the greatest interest to nat- 
uralists, as they presumably developed in Ber- 
muda from related plants formerly existing 
but now mostly extinct there. Of the 61 en- 
demic plants, 11 are flowering plants, 4 are 
ferns and the rest are flowerless species of 
mosses, lichens, fungi and alge. The total 
number of native species known, those that 
have reached Bermuda independently of human 
activities, and have perpetuated themselves, 
including the endemics mentioned above, is as 
follows: flowering plants 146 species; ferns 
and fern allies, 19 species; mosses and hepatics, 
51 species; lichens, 80 species; alge 238 spe- 
cies; fungi at least 175 species. This makes a 
grand total of 709 species. The number of in- 
troduced and completely, or partially, nat- 
uralized species, those which have reached Ber- 
muda through human activities, is about 303. 
It might be added in closing this review that 
all groups of plants are considered in the 
“Flora of Bermuda.” The least satisfactory 
portions of the whole book are those dealing 
with the fungi and the diatoms (Bacillariee). 
The description of the fungi deals with much 
irrelevant matter. It would have been much 
better to have given what is actually known 
about the Bermuda fungi, than to have brought 
in a whole lot of interesting facts about the 
morphology and physiology of this group of 
plants, which can be found in the ordinary 
text-books of morphologic botany, but which 
do not apply especially to the flora of the | 
group of islands under consideration. 
Joun W. HarsHBERGER 
SPECIAL ARTICLES 
THE RYDBERG UNIVERSAL CONSTANT No 
In connection with some work along related 
lines, the author has noticed that Curtis,! in 
his work on the Balmer series of hydrogen, 
1 Proc. Roy. Soc., (A), 90, 605, 1914. 
SCIENCE 
47 
reduced his measured wave-lengths to vacuo in- 
correctly. These wave-lengths were measured 
on the I. A. system, and hence at 15°C., 760 
mm. pressure, while Curtis, for the reduction, 
used Kayser’s Table of Corrections,? which ap- 
plies only to the old Rowland system, founded 
on 20° C., 760 mm. pressure. 
The error thus introduced, for any given 
frequency, is approximately ' (n—1),v where 
nm = index of refraction of air at 0°C., and pv 
is the frequency. For the spectral range of 
the hydrogen Balmer series, as well as of 
all the ordinary helium series (n —1), varies 
by only 2 or 3 per cent. of itself. Thus to this 
approximation the error is proportional to the 
frequency. 
Now the main object of Curtis’s work was 
to test the aceuracy of the Balmer formula, 
and to derive a more accurate value for the 
one undetermined coefficient (the Rydberg 
constant NV,) which occurs in this formula. 
The Balmer formula is 
7 1 1 
»=Na(F-=3); 
where m=—3, 4, 5, ete. Since the error made 
by Curtis is fortunately a constant per cent. of 
the frequency (within 3 per cent.), it will not 
affect the accuracy with which the formula 
does or does not fit the series. It will change 
only the value of N,. Curtis found that he 
must actually use the formula 
1 1 
r= No(a- wee)” 
where # = 69 X 107 and N, = 109,679.22. 
The error as given above amounts (in terms 
of N,) to 0.50 for H,, 0.513 for Hg, and 0.515 
for the remaining lines. The correct value of 
N, is therefore 109,678.705, which agrees very 
closely with the value (109,678.6) determined 
by the author? by direct conversion from the 
Rowland system of the best previous measure- 
ments. With this value, Curtis’s formula will 
fit equally well for all the hydrogen lines ex- 
cept H,. For this line the (obs.-cale.) will be 
— 0.0008 A, instead of + 0.0001 A as given by 
Curtis. Since the error is only in N, and is 
2 Kayser’s ‘‘Handbuch,’’ Vol. 2, p. 514. 
8 Astro. Jour., 32, 114, 1910. 
