544 SECTIONAL TRANSACTIONS.—J. 
nn produce abnormal low and high grades o-10, according to the modifiers 
present. The formula is relatively simple, involving only seven kinds of 
effective gametes, and is of fundamental practical importance since it predicts 
with considerable precision the results of any grade-matings. A scheme of 
family allowances based upon it would maintain minimum number of high- 
grade children necessary for preservation of modern civilisation. 
Division 2. 
Mr. R. J. BARTLETT.—The effect of so-called ‘ constant’ errors in sensory 
comparisons (2.0). 
Further work with geometric series of weighted containers. The ‘ con- 
stant error’ increases in amount as weight increases or decreases from a 
‘datum’ value. This ‘datum’ not a weight or a density but a value 
depending on nature and size of the container. The error has sign and in 
adults appears to be approximately proportional to the cube of the difference 
between the weight lifted and the ‘ datum ’ weight. With children the error 
appears to increase more rapidly. 
With practice, the ‘ constant errors’ for a particular series decrease in 
amount, and possibly would eventually become zero throughout the scale. 
With the series used decrease is more rapid at the heavy end than at the 
light end, and there are indications that after the first few sittings of a subject 
the heavy weights are regressing towards a heavier and shifting datum, while 
the light ones are still referred to the original datum or to one only slightly 
heavier. 
The common experience of subjects that discrimination is easier at the 
heavy than at the light end is supported by the decrease in value of the scatter 
error of the best equal value as the standard weight increases in value, and 
indicates that the true ‘ Weber constant’ (freed from the masking effect of 
the ‘ constant error ’) slowly decreases in amount as the stimulus increases 
in value. 
Mr. M. F. Lowe.—Alterations in blood distribution during mental work 
(2.45). 
In this communication experiments with the Mosso Balance, and also 
with two modifications of it, are described. 
Results from the simple Mosso Balance indicate that the ‘ head end ’ of 
the apparatus becomes lighter during mental work, and not heavier as 
Mosso had stated. Further, it is shown that the conclusions of Ernst Weber 
in regard to the controlling influences of the position (in regard to the axis 
of the balance) of the abdominal organs of the subject must be revised. 
In the first modification the apparatus was arranged so as to rock from 
side to side instead of up and down as the original Mosso Balance had done. 
It is shown that the approach to sleep is accompanied by a gradual depression 
of the left side of the balance, while mental activity is accompanied by a rise 
of the left side. 
In the second modification the balance was constructed upon the gimbal 
principle so that deflections from the up-and-down Mosso motion and from 
the side-to-side motion could be recorded simultaneously. From these 
experiments it is shown that various mental states (e.g. activity, passivity, 
sleep) can be connected with definite combinations of balance movements. 
