The chief obstacle to obtaining accuracy in result is the general conviction among medical practitioners that a different treatment is required, even in fixed morbid conditions, according to the symptoms which may be present. The progress of diseases is never absolutely uniform, and no doubt the occurrence of particular phenomena often require special interference. This secondary treatment of symptoms, however, should never be allowed to interfere with the primary management of the morbid condition; and it is the neglect of this rule which has led to such injurious results in the treatment of many diseases. If, for example, in order to relieve cough in phthisis we give opiates and expectorants, how can we maintain the appetite and improve the tone and digestibility of the stomach, on which the assimilation of food, cod-liver oil, and nutrition essentially depend? Since the publication of my papers and treatise on the Restorative Treatment of Pneumonia I have watched with great interest what has been published by the profession on this subject. The only published series of cases that I am acquainted with is given by Dr. T. N. Borland, of the Boston City Hospital, U.S. He tabulates according to the form I recommended 90 cases of pneumonia, of which he says twelve died—a mortality of one in 7½ cases.[1] Of these, four had phthisis; two were chronic, having been admitted on the eighteenth and twenty-first days of the disease; one was utterly prostrated on admission, and died the following day; one was a case of surgical injury, transferred to the medical wards on the occurrence of fatal pneumonia; and one was a case of typhoid fever—leaving only three fatal cases of true primary acute pneumonia. Of these, one died of cerebrospinal meningitis; a second suddenly, from supposed embolism; and a third, from extensive double pneumonia, with violent delirium. Details of the post-mortem appearances are much desired in these fatal cases. A rigid scrutiny into the true character of these cases therefore shows, instead of a mortality of one in 7½ cases, as is alleged, a real mortality of only one in 27 cases—that is, three deaths in 82 cases.