92 
IN ATRC Tele 
[ NOVEMBER 26, 1903 
division of the Karu, flora, on the other hand, is classed 
as Rhzetic, while the Ecca, or lower Karu, flora is identified 
with some part of the Permo-Carboniferous. The latter 
conclusion, it may be mentioned, is rendered practically 
certain by the recent discovery in Kashmir of Glossopteris 
belew marine Permian strata, as recorded in the report of 
the Geological Survey of India for 1902-3. The occurrence 
in the Ecca beds of Vereeniging of Sigillaria and other 
European Carboniferous types points to a closer connection 
between the South African Glossopteris flora and the Carbon- 
iferous ‘flora of the northern hemisphere than exists 
between the latter and the Glossopteris flora of the Lower 
Gondwanas of India. The Ecca beds of Vereeniging 
appear to be the equivalents of the Karharbari beds of the 
Gondwanas. Finally, although deprecating a_ precise 
identification, Mr. Seward is of opinion that the Witteberg 
flora is probably Carboniferous or Devonian—more likely 
(he latter than the former. 
That so much good work—both strictly scientific and 
economic—should have been accomplished during and so 
soon after a great war is a hopeful sign for the future of 
South African biology. Ro 
EXPERIMENTS ON WHEAT. 
ITHIN the last few years it has gradually been recog- 
nised that, although our wheat-fields produce a large 
bulk of grain, it is, if used alone, unsuitable for the manu- 
facture of the light white bread now generally demanded. 
In consequence, increasing quantities of the harder and 
more suitable wheats grown in Canada, the United States 
and other countries are imported yearly, and the price of 
the inferior home-grown grain has fallen considerably. 
More or less concurrently with this greatly improved 
methods of milling have come into vogue, and the farmer, 
perhaps not unnaturally, associates the two facts, and all 
too frequently blames the miller for his reduced margin of 
profit. A little closer examination of this complicated 
problem shows that the tendency for the last thirty years 
or so has been for the yield per acre of grain to rise, 
and the quality, as estimated by the percentage of gluten 
present, to fall.* 
Now in some way or other, precisely how we do not 
know, the capacity of wheat to yield a strong flour, or its 
““ quality,’’ is bound up in this mysterious mixture of pro- 
teids grouped together as gluten, so that if the blame must 
be apportioned, it rests on those who injudiciously selected 
wheats for cropping power in preference to quality. 
Meanwhile, such fine old varieties as Golden Drop, Red 
Lammas, and Nursery wheats are steadily being driven out 
of cultivation by varieties slightly superior in yield, but far 
poorer in quality. 
The great importance of making the most of our home 
wheat-supply has been insisted on time after time by the 
National Association of British and Irish Millers, and one 
of the methods they have suggested is to raise improved 
strains of these good varieties, either by hybridising or by 
selection. Experiments along these lines have been carried 
out for the last three seasons by the Cambridge University 
Department of Agriculture. In the first place wheats 
known to yield a good quality grain have been crossed 
together with the object of finding more vigorous races 
among the progeny of the hybrids. Further, varieties 
selected from a collection of several hundreds for possess- 
ing such characters as a strong, resilient straw, a short | 
period of maturation, and freedom from various diseases, 
have also been used as parent wheats. 
So far it is early to predict any results of technical value, 
but a number of results of scientific interest have already 
been arrived at in connection with Mendel’s laws of inherit- 
ance. The flowers of wheat being autogamous are 
specially advantageous for such work, as Spielman’s careful 
researches on wheat-breeding, carried out without any 
previous knowledge of Mendel’s work, have shown.  Spiel- 
man has already recognised that lax ears, the lack of awns, 
velvety chaff, and red colour are dominant characters, 
1 The figures are set out in detail in Girard and Lindet’s ‘t Le Froment 
et sa Mouture,” p. ror. (Paris, 1903.) 
NO. 1778, VOL. 69] 
while dense ears, the presence of awns, glabrous chaff and 
white colour are the corresponding recessive characters. 
‘These results have already been amply confirmed. 
Thus from crosses between beardless and bearded wheats 
the resulting hybrids have invariably shown the beardless 
character, while their progeny have consisted of beardless 
and bearded forms in the proportion of three to one. 
Similar results have been obtained on crossing lax and 
dense eared races, rough and smooth chaffed, and red and 
white, though in the last case it has so far been impossible, 
owing to bad ripening, to distinguish clearly enough 
between red and white chaff to establish their proportions. 
At the same time it has been shown that the sharply 
keeled glumes found in Triticum turgidum, e.g. are 
dominant over the glumes with rounded bases occurring 
commonly in the varieties of T. vulgare, that the grey 
colour of glumes and paleze is dominant over red and white, 
that broad leaves are dominant over narrow, and rough 
ones over smooth, that certain groups of bristles on the 
ridges of the stem which distinguish some varieties are 
dominant over the ridges without bristles, and that hollow 
stems are dominant over pithy stems. With regard to 
grain characters, the long and narrow type is dominant 
over the short and round, and the red over white. At the 
same time certain complications have been met with which 
will entail further investigation. Thus the rough-chaffed 
| grey Rivet’s wheat, when crossed with a smooth-chaffed 
white or red wheat, gives hybrids which vary considerably 
both in the roughness and colour of the chaff, some being 
almost glabrous and showing decidedly the red or white 
colour as well as the grey. The same impure dominance 
of the rough chaff and colour is found in the following 
generation. Where other rough-chaffed wheats have been 
made use of in the place of Rivet wheat though this 
character has been purely dominant. 
Further, particularly among the progeny of the hybrids, 
there is a marked tendency for the various characters to 
become intensified. Medium lax, for instance, becomes very 
lax, the grey colour becomes almost black, ard the red a 
deep brown. At the same time, unexpected forms appear 
in this generation showing characters unrepresented in 
either parent. The commonest of these, so far, has been 
a spelt-like wheat with peculiarly lax ears, thick glumes, 
and the typically closed spikelets of T. spelta. Many of 
these exceptional forms are sterile—probably owing to im- 
perfectly developed pollen. 
These botanical characteristics are, however, of little 
importance technically, the farmer and miller being con- 
cerned chiefly with the quality, yield, hardiness, time of 
ripening, susceptibility to disease, &c., characteristics, at 
present practically unexamined, which one might term 
“* constitutional.”’ 
The quality of the grain can, to a certain extent, be 
judged by the hardness and translucency of its endosperm, 
the poor starchy grain being soft and opaque. Accepting 
this as a guide, then, good quality is a dominant character, 
at all events so far as an examination of the first gener- 
ation of the hybrids goes. The late ripening habit is also 
dominant over the early ripening habit. As an example, 
T. Polonicum, ripening early in August when sown about 
the middle of March, was crossed with Rivet wheat ripen- 
ing late in August when autumn sown. The hybrid grains 
were sown on March 15, and produced plants which 
| ripened their grain about the middle of September—simul- 
taneously with Rivet wheat sown on the same date. 
Experiments on the susceptibility to disease are also being 
carried out. This point is being investigated both with 
rusts and mildew, the two serious wheat diseases, inasmuch 
as they are untreatable. For the purpose of the experiment, 
in 1g0r Michigan Bronze and a wheat with the Michigan 
Bronze strain in it, viz. Red King, both liable to rust, 
were crossed with Rivet wheat, which is practically 
immune. Reciprocal crosses were made in each case. The 
following year che hybrids were the most badly rusted 
plants among the experimental plots, and there was nothing 
to choose between the plants with Rivet wheat as male or 
female. Incidentally, then, it might appear to anyone who 
accepted Eriksson’s views that in the case of Rivet wheat 
Q x Red King or Michigan Bronze ¢, the so-called 
“mycoplasma ” had reached the hybrid grain by way of 
