SOME PROBLEMS OF ATMOSPHERIC CHEMISTRY 
gives an insight into the distribution of the sulfate up 
to 50 km leeward of intense sources of gaseous effluents. 
Exact instantaneous measurements can, of course, be 
made only with the apparatus discussed above in the 
paragraph on the determination of iodine. For this 
purpose from 0.05 to 20 m° of air (depending on the con- 
centration) are washed in aqueous solutions of alkali 
hydroxides or carbonates, a process which requires 
5-120 min. Subsequent analysis requires about 60 min. 
Hydrogen sulfide is determined, according to Quit- 
.mann [45], by absorption in 2-3 cc of acetic cadmium 
acetate solution, with the smallest Cauer apparatus. 
For this measurement, 300-500 | of air are washed in 
about 60 min. The end determination is performed with 
KT solution in approximately 30 min. 
Nitric acid is determined in the same way as sulfuric 
acid, in dilute aqueous solutions of alkali hydroxides, 
using the same equipment. After reduction with metallic 
zinc, the nitrate is determined as nitrite with Griess’s 
reagent [47]. The time required corresponds to that 
needed for the determination of H»SOu. 
Ozone is determined, according to Cauer [6], with 
the aid of the intermediate-size apparatus described 
previously, which is provided with an absorption tube 
(circulation capacity 0.6 m* hr). The classical method 
of detection, based on reaction with KJ in a neutral 
aqueous solution, is used. From 100 to 300 | of air are 
washed within 10-80 min in 10 cc of 0.2 to 0.6 per cent 
aqueous NaC,H302 solution (pH 6.7-7.1) which con- 
tains 50 ug iodine ions as KJ. The iodide is oxidized to 
In by O3 and the Ip is removed by the air stream. There- 
fore, in the end determination the residual nonoxidized 
iodide is analyzed according to the method of von 
Fellenberg [25], and the difference is computed in terms 
of O3. For experienced chemists, this analysis requires 
20 min, but necessitates great care, and is difficult for 
nonchemists. By working in shifts, four trained opera- 
tors can make two or even three parallel series of hourly 
measurements for several weeks at a stretch. The 
method is inapplicable if the air contains relatively 
large quantities of nitrite and of free halogens which, 
because of their acidic character, nullify the buffer 
effect. Unpublished experiments, made on the North 
Sea, have shown that even at a high natural content 
of total Cl, the buffer effect is not impaired, because 
the dilute free chlorine present is evidently converted 
very rapidly into chlorine dioxide. A corresponding 
absence of significant errors has also been noted in the 
presence of large quantities of natural nitrite [13]. 
A similar method, published later by V. Regener [49], 
must be considered a close parallel to the foregoing 
technique because of the addition of the buffer effect 
contributed by Ehmert. Unfortunately, because of dif- 
ficulties inherent in the apparatus, this procedure is 
even more awkward to carry out than Cauer’s method. 
Simplification of the method for O; determination 
appears to be an important problem. In the author’s 
opinion, the approach to this problem is hardly to be 
found in deep cooling and absorption in silica gel (be- 
cause of increased possibility of error), but perhaps in 
1127 
the oxidizing effect of O; on aldehydes, as attempted by 
Briner and Perrottet [4]. The approach used by Dirnagl 
[22] and Curry [19] for developing automatic recording 
of the KJ method has much to recommend it. This 
apparatus is a development of the automatic recorder 
for the determination of HS, according to Kraus [32]. 
According to their technique, a large unmeasured quan- 
tity of air is blown mechanically on a filter paper soaked 
with a neutral AJ solution. At stipulated time intervals 
a fresh portion of the filter paper is exposed by clock- 
work. The resulting coloration can be calibrated by the 
exact methods discussed previously. An improvement of 
this method and a reduction of its cost would offer new 
possibilities for more extensive comparative investiga- 
tions. In this respect, to be sure, technical deficiencies 
cannot be justified by assuming that an additional 
oxidizing factor, such as O,, might be present, or even 
that the total oxidation value is being measured [22]. 
The latter is certainly unlikely, for NO: and ClO, 
oxidize in acidic solution at the pH value used, but by 
no means completely. 
The total oxidation value is determined, according to 
Cauer [6], in the same way as is O;; not in neutral 
solution, but in H»SO, solution (0.05 per cent) having 
a pH below 2.8. The oxidized J» is calculated in terms 
of active O, in which two iodine atoms correspond to one 
oxygen atom. The time required for sampling and analy- 
sis is the same as for O3. 
The total reduction value, according to Quitmann 
[44], is determined with the aid of the smallest of the 
Cauer absorption tubes, by washing 20-50 | of air in 
3 ce of a concentrated K2Cr.0,-H2SO. solution for 
15-30 min. The final determination consists of the 
addition of AJ and titration of the liberated iodine with 
thiosulfate. The difference between this iodine value 
and the one which would have been determined if no 
chromate had been reduced by substances in the air 
corresponds to the reduction effect of the air and can 
again be computed (according to the formula: 2 iodine 
= 1 oxygen) in terms of the active oxygen missing for 
the oxidation of the reducing substance (see Table III). 
In this way the total oxidation values can be compared 
directly with the reduction values. If the reduction 
value is divided by the oxidation value, the quotient 
represents the Redox value of the air. During World 
War II a simple, rapid method for determining the re- 
duction value, designed for hygienic purposes, was 
developed by the author. For meteorological purposes, 
this method, as yet unpublished, might also be appli- 
cable to air that is not too humid. 
Condensation Method for Determining Traces of 
Substances Which Form Condensation Nuclei. Con- 
densation nuclei are obtained, according to Quitmann 
and Cauer [47], or according to H. Cauer and G. Cauer 
[15], by cooling the highly polished metal surfaces of a 
Cauer condensation sphere below the dew point but not 
to the freezing point. The water in the thin layer of air 
very close to the metallic surface is rapidly attracted by 
chemical substances suitable as nuclei, and droplets are 
