SEPTEMBER 8, 1899. ] 
The method of investigation which is most 
often employed is that of the statistical inquiry, 
but this is supplemented by the methods of per- 
sonal questioning, observation of the sleeper’s 
movements and experimental stimulation. 
Practically all the methods by which dreams 
have been studied are therefore employed, ex- 
cept that of ‘direct observation,’ by which the 
dreamer, immediately after waking, records his 
own dreams and notes their vividness, their re- 
lation to waking experience and other important 
features. De Sanctis justly criticises this method, 
on the ground that the, intention of studying 
one’s dreams is itself an artificial condition, pre- 
disposing the subject to dreams of unusual fre- 
quency and of unnatural content ; but the diffi- 
culty, which undoubtedly exists, he distinctly 
overstates, for individuals differ greatly in their 
ability to preserve a normal disposition under 
artificial conditions. The writer of this notice, 
for example, observed her own dreams, after 
the method already described, for nearly two 
months. The work of recording the dreams 
and their conditions was performed with me- 
chanical, and, so far as possible, with unreflec- 
tive, accuracy ; and the study of the records was 
not undertaken until the completion of the ob- 
servations. The result in this case was the 
record of about two hundred dreams, which 
were certainly very closely representative of 
the ordinary dream-life of the observer and 
noticeably destitute of unusual or abnormal 
features. 
On the other hand, De Sanctis does not suffi- 
ciently emphasize the disadvantages of the sta- 
tistical method which lies at the basis of the 
greater part of his conclusions. The extreme 
liability, varying as it does with individuals, to 
forget one’s dreams, throws grave doubt upon 
the answers of people, untrained in introspec- 
tion, to questions about the frequency and the 
vividness of dreams, the emotional nature and 
the connection with waking experience. 
The uncertainty of the inference from bodily 
motions to the accompanying facts of conscious- 
ness, when these cannot be tested by the wak- 
ing memory of the sleeper, seriously affects the 
conclusions of the chapter on the dreaming of 
animals. The discussion of children’s dreams, 
on the other hand, is illuminating and suggestive 
SCIENCE. 335 
in so far as it is based upon the author’s per- 
sonal study of the dream-life of his own chil- 
dren. He concludes that children begin to 
recall their dreams at four or five years of age 
and he identifies this period with the epoch of 
the distinct consciousness of self ; but he con- 
cludes that children actually dream before the 
years when they recall their dreams, from the 
fact that characteristic movements in sleep, 
such as laughter and irregular breathing, which 
are later proved to accompany dreams, do ac- 
tually occur before the fourth year. 
Only twenty subjects of advanced age were 
questioned about their dreams, and these con- 
firmed the ordinary statements concerning the 
infrequency and the colorlessness of the dreams 
of the aged. The fact that only one of these 
twenty reviews, in her dreams, the life of her 
youth, confirms the results of experimental 
studies in waking association, and shows that 
old people differ, like younger ones, in their 
tendencies to recall the distinct periods of their 
lives; some of them, waking and sleeping, oc- 
cupy themselves mainly with the past, but 
others live a life full of present issues. 
The chapter on the dreams of adults includes 
summaries of earlier work on the same line and 
the results of statistical inquiry comprising 
answers from one hundred and sixty-five men 
and from fifty-five women. These figures are 
too disproportionate in themselves to permit 
the comparison, which De Sanctis proceeds to 
make, between the dreams of men and of wo- 
men. His conclusions, however, while numeri- 
cally very different from those of Heerwagen, 
are of the same general nature; he finds (p. 
135) that women’s dreams are more frequent, 
more vivid and better remembered than those 
of men. Likeall other investigators he shows 
also the close connection of dreams with wak- 
ing experiences. 
The discussion of the inquiry which follows— 
statistical and personal—into the dreams of the 
mentally deranged is itself too condensed to be 
readily summarized. Imbeciles 22a epileptics 
(except those slightly affected) are found to 
dream infrequently ; hysteric patients, on the 
other hand, and paranoiac subjects are set down 
as constant dreamers. 
The last of these comparative studies, that 
