74 Rhodora [Max 
eighteen months for the germination of the seed and from eight to 
ten years to secure mature fruit from the first generation. ‘Thus one 
would need the equipment and backing of an arboretum to carry on 
such work on a scale sufficiently extended to secure definite results. 
So far as known, the genus produces plenty of good seed, and the 
plantations at Biltmore, the Arnold Arboretum, and my own, at the 
New York Botanical Garden, have shown that the forms of the genus 
reproduce themselves surprisingly true from seed. Yet, further 
experiments are greatly to be desired, in which a// the seed from a 
single tree are retained and raised to maturity. When this is tried, 
it will be surprising indeed if Crataegus differs from other members 
of the Rose Family, which are so notoriously susceptible to crossing. 
'The genus is so variable that it is impossible to find characters 
which will hold absolutely true, either between species or even sections; 
then again, characters which seem to hold good in one section are 
useless in the next. Thus in the Molles and Coccineae certain species 
have cordate leaves on their vegetative shoots, but in the Tenuijoliae 
and Pruinosae both cordate and wedge-shaped leaves occur on the 
same plant. 
The best time to study Crataegi is when they are in mature fruit. 
Often it is impossible to name dried flowering material closer than to 
one of two groups, while mature fruiting material, even in the dried 
state, would easily settle the species. This case holds particularly 
true between the T'enuifoliae and Pruinosae, the Coccineae and Molles, 
the Macracanthae and Anomalae. 
Personally I am fast losing confidence in the number of stamens as 
a reliable diagnostic character. It is true that in a general way the 
flowers of many species seem to have prevailingly 5-10 stamens or 
10—20 stamens, but both 10-stamened and 20-stamened flowers some- 
times are found on the same tree; and material is fast accumulating 
which seems to show that for every 10-stamened form somewhere a 
similar form will occur with 20 stamens, and the reverse. 
In a general way color of anthers seems to be correlated with color 
of the immature foliage; thus yellow anthers go with yellow-green 
leaves, pink anthers with bronze-green leaves; and the different 
sections have their own particular color. For instance, the only 
constant difference between the Coccineae and the Molles is this color 
difference, all other distinctions breaking down at some point; but 
even this character must be employed with caution, particularly if 
