The question of whether the secondary infection occurred before or after the patient was admitted to the hospital was in many instances impossible of exact determination. In order to arrive at a fair opinion, a third doubtful class is added. For the purpose of this classification I have regarded any case developing measles before the ninth day of admission as having contracted his infection before admission and after the eighteenth day as having contracted it in the hospital. Those developing the infection between the ninth and eighteenth days have been considered as doubtful cases. They may or may not have been infected in the hospital. While this is not an absolutely accurate method of determining the truth of the matter it is believed to approximate the truth as nearly as possible. Undoubtedly there are cases where the symptoms of measles have been delayed for more than 18 days after exposure, and possibly there are some authenticated cases where they have been noticed before the ninth day, but the concensus of opinion seems to be that the general average of 14 days is but seldom subject to more pronounced variations than these. I have dated the onset of measles from the appearance of the catarrhal symptoms.

Adopting the same general principles in regard to incubation periods of the other disease, I have diagnosed as diphtheria only those

cases which were both clinically and bacteriologically such. Diphtheria carriers are not included in this list. Any case thus developing diphtheria before the third day's residence in hospital was regarded as having contracted the disease before admission. From the third to the fifth day it has been regarded as doubtful, whereas after the fifth day it has been considered as a case of infection occurring in the hospital. Scarlet fever has been diagnosed as such from the first appearance of the rash. All cases occurring before the third day as having contracted the disease before admission. From the third to the eighth day it has been regarded as doubtful, and after the eighth day as cases of hospital infection.

Chickenpox was regarded as having been contracted in the hospital if it occurred before the sixteenth day. The fourteenth to sixteenth day was considered doubtful, while under 14 days it was regarded as occurring before admission.

Mumps was considered doubtful if occurring between the seventeenth and twenty-first days; before the seventeenth day as before admission, and after the twenty-first day as a hospital infection.

Whooping cough was counted as such from the first characteristic whoop and 16 days was assumed to be its incubation period. Any case developing under a 16 days' stay in hospital was considered as having contracted the disease before admission and any after that as a case of hospital infection.

German measles contracted under 10 days was considered as an outside infection; 21 days and after as a hospital infection, and between those times as doubtful.

By the foregoing tables (pages 110 to 114) we see that 84 were exposed to the second disease prior to admission, leaving 73 who contracted the disease either after admission or at a time when the incubation period would place them in the doubtful column. If we consider the whole 73 as having contracted the second infection subsequent to their admission to hospital this will give us a percentage of 2.7 cross-infection to be charged against faulty technique or unavoidable exposure.