100 '^- Lindhard. 



in the direction of breathing out more quickly or more explosively. 

 The number 458 is the average of 6 analyses; it contains a good 

 deal of meaning, taken with the remaining results; but the single 

 determinations vary from a little over 4 to a little over 5, thus 

 much more than the whole series from day to day. The last value 

 is the average of 2 analyses, 46 and 44. It is noted for the last 

 sample, as for some of the samples on the previous day, that the 

 expiration was very short and explosive. I am thus of the same 

 opinion as Loewy, that the expiration should take place quietly 

 and controlled; otherwise, vortices will easily be caused in the 

 expiration-air, which mix this with the air in the "dead space" and 

 thus affect the results. 



If w^e carefully observe this rule, we should obtain values from 

 Haldane's "inspiration-sample", which agree with those calculated 

 for the average of the expiration. Haldane and Priestley's second 

 sample is taken, as mentioned, from the reserve-air and will there- 

 fore, as my results also show, always give too high values. 



As the standard deviation in my series of the averages is only 

 019, I venture to conclude, that I had complete control over the 

 technicalities and presuppositions connected with the taking of the 

 samples. 



There seems to be a certain connection between the height of the 

 barometer and the alveolar CO., tension, the variations of the latter 

 proceeding on the whole parallel with the changes in the air-pressure. 

 Looking at the matter broadly, when the barometer < 762, the CO^ 

 tension = 34- 1, whilst when the barometer is > 762 the CO 2 ten- 

 sion = 34-3. Any influence from the temperature cannot be detected. 



The production of carbonic acid varies inconsiderably 

 from day to day. Experiment No. 132 is the only one that differs 

 in this as in several other regards; the amount of carbonic acid 

 given off on that day was as much greater than the next highest 

 value, as the difference between the latter and the lowest value of 

 the series. Calculating the limits for the series from the formula 

 fxk, where к is found from the integration table for the factor 



(n being as usual the number of variants in the series), this 

 4n ^ ° 



experiment falls outside the series; I have included it, however, as 



it does not exceed more than 3pt from the average in any of the 



columns calculated. 



For the whole series we have: average = 194 4:; 088, /л = 435 

 or 2-2^lo of A. Omitting Exper. 132 we have: average = 193 + 

 0-57, fx = 2-66 or 1-40/0 of A. 



The bodily weight remained practically unchanged during the 



