First, carrot infusion, two months old; second, calf jelly; third, ascitic serum, collected antiseptically, three weeks old, and perfectly clear.

Pericardium experiments.-Under the spray, and with strict antiseptic precautions, a piece of the pericardium was removed, little pieces of lymph were scraped off, and a flask of serum was inoculated and placed in an incubator at 90° F. The upper part of the fluid remained clear, but a deposit formed at the bottom. The germs were anaerobic. Examined microscopically, they were found to contain the same micro-bacilli above referred to, without nuclei. Similarly two other sterilized flasks of serum were inoculated from the flask with the same results.

Microscopic section of heart.-The section of the pericardium showed wavy layers of fibrin and lymph more or less organized. Strewn in the meshes were small clusters of micro-bacilli.

Wilson concludes that the pericarditis was undoubtedly associated with a micro-bacillus, which was the primary germ or parasite.

Dr. T. M. Rotch of the Harvard Medical School has kindly recapitulated for us his views on the percussion outline of pericardial effusion and the selection-site for paracentesis pericardii, making his deductions from experiments upon sixteen infants; we take great pleasure in incorporating them in this little work:

"The following remarks concerning pericardial effusion are a brief résumé of the observations which I have made on this subject since publishing my article in 1878. These observations, which are mostly confirmatory of the points suggested in the original article, have been made at the Boston City Hospital from its large number of acute cases, amounting to nearly three thousand ward-patients in a year of all ages, from the Children's Hospital, Infant Hospital and the Boston Dispensary children's clinic, the clinical observations also being tested in various ways in the anatomical and physiological laboratories of the Harvard Medical School.

"Owing to the acknowledged latency of the symptoms in

many cases of acute pericarditis in infancy and childhood,* the diagnosis of effusion must often rest on the physical signs, and of these signs, as I have already stated,† percussion is the only one that can in the majority of cases be relied upon at the stage of the disease when it is especially important to know whether an effusion is present or not for purposes of differentiation from dilated heart, with the question of tapping to be decided upon. Again, this physical sign of percussion can only be of value when applied after precise anatomical relations have been established and physiological and mechanical laws recognized; and this at once brings us to the anatomical consideration of the subject.

"In the various articles written on pericardial effusion the authors are found to copy each other year after year, expounding, repeating, and perpetuating views which have never been distinctly proved and are often manifestly erroneous, their lines of percussion not being exactly stated and their diagrams, from improper methods of experimenting, being open to criticism, as representing hypothetical conditions which do not exist in the human subject.

"Hughes states that infantile pericarditis occurs more frequently than the adult disease, and the younger the child the more chance there is for the appearance of the disease.§ Keating and Edwards speak of the association of endocarditis and pericarditis as a frequent concomitant, and of the dilatation of the cardiac cavities being caused, to a great extent, by adhesions in the adult, and by the myocardium being affected in the child. Keating and Edwards have also given a most valuable description of the anatomy, physiology, and general topography of the heart and pericardium.¶

"Assuming that the observations of these gentlemen are correct, they have a significant bearing on what I consider of the first importance for a proper appreciation of the subject—

Keating and Edwards, Archiv. Pediat., April, 1887, p. 209. †Transactions Mass. Med. Soc. 1878.

A notable example of this statement can be seen in the pericardial-effusion plates in Reynolds' System of Medicine, copied by Roberts and others. Archiv. Pediat., April, 1887, p. 204.

Ibid., February, 1887, p. 65.

|| Ibid., p. 208.

namely, the methods of the anatomical investigations. The older the child and the nearer to adult age that it approaches, the greater is the chance that adhesions from some slight and perhaps almost overlooked inflammation of parts or organs adjacent to the pericardium have taken place, and this, in connection with the natural absence of adhesions mentioned by Keating and Edwards, emphasizes the fact that the primary experiments regarding the shape of a distended pericardium should be made on the infant or young child, where the result is more likely to show a correct figure on percussion, both clinically and in the laboratory.

"The number of clinical observations on infants is not yet large enough to provide us with sufficient data, but the experiments from which my diagrams are drawn were made on sixteen infants, in none of whom did adhesions exist, while in the case of an adult at the city hospital-where, although the pericardium was evidently much distended with fluid, the percussion failed to show absolute dulness to the right of the sternum-the autopsy revealed adhesions of the middle lobe of the lung holding it tightly to the right edge of the sternum in the 5th interspace, with the effusion behind it, thus giving resonance where the uncomplicated case would have given dulness; hence it is evident that we must first study and acquire a knowledge of the uncomplicated cases before we are prepared to elucidate cases complicated by adhesions. And yet there is a strong probability that many of the clinical observations made on adults by various competent clinical observers, and established by them as rules for diagnosis, are, from the presence of adhesions-sometimes in one place and sometimes in another-rendered of little practical value, as shown by the difficulty experienced in making a diagnosis by these rules in new cases, and even delaying the final solution of the problem by, through not knowing where or whether, and hence not stating where or whether, adhesions existed, perpetuating erroneous views.

"Assuming, then, for the present that the infant's (first) and the child's (second) pericardium is most likely to be the best for studying the uncomplicated pericardial-effusion outlines, and allowing that experiments made as directed in my

article in 1878 are correct until they are disproved-which so far has not been done-the deductions which can be made from the results of these experiments are that the fluid accumulates at the bottom of the pericardial sac, where it collects on either side of the arched diaphragm like a saddle,

DIAGRAM I.-Small amount of liquid introduced into sac (Rotch).

A, The portion of the area of absolute dulness which is still caused by the physiological dulness of the heart.

B. Liver.

B, That portion of the liver which is
covered by the right lung.

C, Lung.

A+ D, Area of percussion dulness found when the effusion is small.

S, Sternum.

D, Effusion.

extending to the right and to the left of the sternum; that the layer of fluid is a little the thickest in the left diaphragmatic depression, almost as thick in the right, and then that it gradually grows thinner as it ascends in the vertical line, and in a small effusion leaves a portion of the heart uncov

ered in the region of the 4th left costal cartilage, as seen in Diagram II.

"The lungs gradually recede before the fluid, as seen in Diagram I., and the absolute dulness is found to the right of the sternum. Of this area of dulness, that portion which

DIAGRAM II.-Represents Diagram I. with the lungs removed (Rotch).

is found in the 5th right interspace is of diagnostic value, first, because, unless there is consolidation of the middle lobe or a right pleural effusion, absolute dulness is extremely rare in this interspace, as proved by my having had hundreds of infants and children percussed in this interspace, with a resulting resonance; and second, because this interspace is not