The extraordinarily large amount of statistical material which exists (P. Friedrich, F. Schmid, German collective investigations) is only of partial value. The usual percentage calculations are unreliable ; in the first place, on account of the enormous number of mild influenza cases which do not come under the notice of the physicians, and also, as experience has proved, because numerous grippal pneumonias were erroneously classed as genuine. Unquestionably, more pneumonia of a slight and mild character occurs than is diagnosed clinically (latent pneumonia).

The frequency of inflammation of the lung in earlier epidemics has been estimated at from 5 to 10 per cent. (Biermer). In the official report in Bavaria the results vary in the different medical districts between 0.4 and 21 per cent., the average being 5 per cent. In the report from the Kingdom of Saxony the number varies between 0.5 and 15 per cent., the average being 7 per cent.; in Wiirtemberg, the same, 7 per cent. The German collective report (Litten) makes a careful investigation of this subject; it gives an average of from 6 to 8 per cent., and shows, apart from some large variations, a satisfactory agreement with the average pneumonia frequency in the separate German states and Prussian provinces. F. Schmid, in the Swiss official report, in a total of 3230 influenza cases, reports 93 pneumonias, or 2.8 per cent.

A well classified outpatient material has unquestioned advantages. Fleischer, in Erlangen, notes that among 543 influenza patients there were only 3 cases of pneumonia, which would point to an unusually mild epidemic. On the other hand, Colley, in the Polyclinic at Greifs wald, among 187 influenza patients saw 14 pneumonias, or 7 per cent. Gmeiner, physician to a glass factory in Bohemia, notes, in his carefully prepared statistics, that among 403 influenza patients, 90 pneumonias occurred, or 22 per cent., and that among these, 64 per cent, affected men, 14 per cent., women, and 21 per cent., children. According to our statistics, influenza pneumonia was more common among men than among women, a circumstance worthy of note, since genuine croupous pneumonia is certainly more common among men than among women.

According to the statistics of von Coler, the frequency of pneumonia in the Prussian army was extremely low. Among 34,556 patients, only 219 pneumonais occurred (0.6 per cent.), and in the complete army report (55,263 cases) 534 pneumonias (1 per cent.), and 175 cases of acute pleurisy (0.3 per cent.), are noted. In a girls' boarding school in London, C. Bristowe reports, among 177 influenza patients, only 3 cases of pneumonia (1.7 per cent.). Hospital statistics, as a whole, naturally give a greater pneumonia frequency. In the Municipal Hospital at

Cologne it amounted to 24 per cent.; in the Hotel Dieu, in the epidemic of 1837 (Copland), to 22 per cent.; in Friedrichshain (Berlin), to 22 per cent.; in Hamburg, to 17 per cent.; in Nuremberg, to 12 per cent.; in Freiburg (Baumler), to 11.8 per cent.; in Riga (Krannhals), to 4 per cent.; in Munich, to 4 per cent.; in Leipsic (Jacob's Hospital), to 4 per cent.; in Magdeburg, to 3 per cent.

That pneumonia is the principal cause of death from influenza we have already noted in the section on Mortality. Of the numerous more detailed statistics in reference to this subject those of Carlsen, in Denmark, are especially noteworthy: Among 502 fatal cases of influenza, 273 were due to complications with pneumonia; 88 to other pulmonary and pleural affections; 49 to pulmonary phthisis; 33 to cerebral disease, while the remaining 59 are registered as simple influenza.

Even if we admit that, as regards the frequency of influenza pneumonia in the statistics of physicians, outpatient departments, and hospitals, chance has played a great part,-and to this cause the enormous variations are partly due,-we must, nevertheless, conclude (compare P. Friedrich's communications) that, as regards the frequency of grip pneumonia, locality unquestionably was an important factor; from several places it is even noted that pneumonia as a complication of influenza did not occur at all. If we wish to draw a conclusion from this statistically proved fact, we must admit that the influenza pneumonia due to mixed infection depended on temporary and local conditions. We content ourselves with the mere allusion to this important subject.

Numerous observers have expressed the opinion that pneumonia in the beginning of the epidemic of 1889 was rare, but became more frequent in the further course of influenza and toward the end of the influenza period. This increase may be apparent because, at the beginning of the epidemic, the complication was only partially or hardly at all recognized.

From our own statistical materials it would appear that the curve of frequency of influenza pneumonia ran parallel with that of the morbidity curve of influenza. There is an entire accord among German and English authors (Wutzdorff, Parsons, and others) that influenza pneumonia occurred much more frequently in the later epidemics than in the pandemic of 1889-1890. (Compare section on Mortality.)

As regards the danger to life from influenza pneumonia, we agree entirely with those numerous observers who consider this form of pneumonia to be more dangerous to life and its mortality considerably higher than that of the "genuine croupous pneumonia." The statistics, it is true, are not concordant. Occasionally they are even favorable; but they do not give a true picture, for the reason that numerous cases of pneumonia which ended fatally in the influenza period were erroneously counted as geniune pneumonias, and, therefore, were not counted as deaths due to influenza. The enormous statistical increase of the general mortality from pneumonia in the influenza period proves this. (Compare p. 568.)

The German collective report gave a mortality from influenza pneumonia of 17 per cent, (varying between 15 and 26 per cent,), a figure which is certainly not greater than the average mortality of genuine croupous pneumonia. In this question, however, hospital observations have a decided advantage, if merely on account of the control of diagnosis by autopsy. In the pandemic of 1889-1890 in the Municipal Hospital at Cologne, among 105 influenza pneumonias 32 terminated fatally, or 30 per cent. Krannhals, in his carefully prepared statistics for Riga in December, 1889, among the influenza pneumonias, records a mortality of 43.9 per cent. The pneumonia mortality in Boston varied in patients at different ages, and was between 29 and 45 per cent. (Mason).