NATURE 



[August io, 191 i 



over, that it is necessary to take into account the absorp- 

 tion, and therefore his formula contains an exponential 

 term to allow for this. 



A perhaps still more important discovery mentioned in 

 the paper is that of Dr. Louis Cohen, that if the reduc- 

 tion in the strength of the received current due to absorp- 

 tion be written e _Arf . then A is inversely proportional to 

 the square root of the wave-length within the limits of 

 accuracy of the experiments. 



Mr. Austin again checks Mr. Marconi's statement that 

 the received signals are proportional to the height of the 

 two antenna;, and adding to this the observation that they 

 also vary inversely as the wave-length, he obtains a com- 

 plete formula giving the received current Ir in terms of 

 the transmitted current I s the heights of the two antennae 

 hji 2 , the wave-length X, and the distance d. The 

 formula is 



Ir=05 



IsV's 



\ I 



Where the currents are in amperes and the lengths in 

 kilometres, the two constants 4-25 and 00015 may depend 

 on the conditions under which the experiments were made, 

 and it will be of great interest if other wireless workers 

 will check the formula against their results and see how 

 closely it is applicable. It must not be expected that this 

 formula will be closely confirmed by every observation. 

 Mr. Austin's own observations show that this is not the 

 case. In spite of the wide range of values he has dealt 

 with, the observations do appear to group themselves 

 round the smooth curves given by his formula. 



The formula refers only to flat-topped aerials and to 

 general day conditions. Mr. Austin remarks that the night 

 signals are entirely irregular, being, in general, stronger 

 than the day signals, and this he assumes is due to there 

 being much less absorption at night, that is to say, the 

 inverse distance law is then more nearly obeyed, even for 

 very long distances. 



It is perhaps of interest to compare Mr. Austin's formula 

 with the measurements in the Monarch tests. Putting 

 the data into Mr. Austin's formula, and taking the wave- 

 length at 250 metres, which was approximately the case, 

 the received current at a distance of 60 miles given by the 

 formula is 590 microamperes, whereas it was actually only 

 about 50 microamperes. It is evident, therefore, that the 

 constant 4-25 is too large for this case. One reason for 

 this may be the great difference in the type of aerial used, 

 Mr. Austin's formula applying to a flat-topped aerial, 

 whereas a straight aerial was used for the Holyhead- 

 Howth experiments ; another reason, the Howth aerial had 

 a higher resistance. 



The absorption coefficient, however, seems to fall in 

 very well with the Monarch experiments, neglecting the 

 short distances, which are irregular. Taking the slope of 

 the curves for the Monarch crossing from Howth to Holy- 

 head, the absorption is rather less than that given by Mr. 

 Austin, but the slope of the curve for the Monarch return- 

 ing from Holyhead to Howth indicates a slightly greater 

 absorption. 



A number of tables are given in the paper to facilitate 

 the use of the formula in practice. These tables show 

 how extremely important it is to use a long wave-length 

 for long distances ; fur instance, for transmission over a 

 distance of 2000 miles, with a wave-length of 1000 metres 

 and two flat-topped aerials 450 feet high, 490 amperes is 

 required in the transmitter, whereas at 6000 metres only 

 105 amperes is necessary. There are still, however, many 

 obscure points in the long-distance transmission which Mr. 

 Austin's formula does not account for; for instance, Mr. 

 Marconi pointed out at the Royal Institution a short time 

 back that there were two minima near sunset and sun- 

 rise in the curve representing the strength of the received 

 signals across the Atlantic, and also two maxima. Can 

 this be accounted for purely by variation in the absorp- 

 tion coefficient, and, if so, does the absorption coefficient 

 during the minima bear the same relationship to the wave- 

 length as that given in Mr. Austin's formula? Do the 

 two maxima correspond to practically no absorption, or 

 are they higher values than would be obtained if no 

 absorption existed as if waves were concentrated, as Mr. 

 Austin seems to consider possible? 



NO. 2l8o, VOL. 87] 



Whether the formula turns out to be strictly right 1 

 not, it should form a good basis on which to compar 

 different wireless systems, and it constitutes a real advanc 

 in the published knowledge of long-distance radic 

 telegraphy. W. Duddell. 



EXPERIMENTS ON AERIAL PROPELLERS." 



AN article in the April Bulletin de la Sociiti d'En\ 

 couragcmenl deals with some experiments on aeria- 

 propellers made by MM. Legrand and Gaudart, with th. 

 aid of a grant from the society. The greater part of th 

 article is a discussion on the methods adopted by othe 

 experimenters for expressing their results. M. Legrant 

 objects to the three coefficients usually adopted in express 

 ing the results of propeller experiments, namely, "pitch,' 

 " fraction of pitch in each blade," and " percentage slip.' 

 He objects to the use of " constructional pitch " (which i: 

 usually taken as the pitch of the pressure face chords), a> 

 it is not constant for all parts of the blade in moderr 

 propellers. He also objects to the use of the pitch corre 

 sponding to no thrust, as this is not constant for al 

 speeds ; but in our opinion this latter is constant enougl 

 for all practical purposes. 



M. Legrand 's objection to the use of the coefficient 

 " fraction of pitch in each blade " is that it is not definite 

 for a given propeller ; as, in modern propellers, it is not 

 lh.' same for all co-axial, cylindrical sections of the blades. 

 This objection, however, is entirely overcome by using 

 "disc area ratio," which is equivalent to "fraction ol 

 pitch," and is also absolutely definite for any given pro 

 peller. The objection advanced against tin- use ol 

 " percentage slip " is that the pitch not being definite, 01 

 the same for all parts of the blade, the slip is also 

 indefinite. 



Iiiii iency curves by G^ber and Dorand are quoted, in 

 which efficiencies at constant rotational speeds are plotted 

 against transIation.il speed. If, however, efficiencies at 

 constant rotational speeds are plotted against percentage 

 slip, and the pitch used in the reduction of the experi- 

 mental results be stated — the percentage slip being equal 



/ pitch X revs. — translational speed \ .. . ,., 



to too . , ) — it is readilv 



\ pitch X revs. / 



seen that the two sets of curves are equivalent and derivable 

 from each other. Also, plotting against percentage slip 

 has the advantage that it brings all the efficiency curves 

 close together. 



It is generallv admitted that the indefiniteness of the 

 pitch of a propeller is a disadvantage; hut ir seems, as vet. 

 lo be the best " coefficient " that can be used to give a 

 general idea of the type of a given propeller. M. Legrand 

 does ii>i give any substitute for "pitch," and, in connec- 

 tion with his own experiments, differentiates between a 

 propeller with a big pitch and one with a small pitch. 



The experiments were carried out on full-size propellers, 

 mounted on actual aeroplanes and driven by a 50 h.p- 

 Gnome engine. The thrust was registered during th'' whole 

 f! ; ght on an autographic diagram from a Richards dynamo- 

 meter, working in conjunction with a flexible mounting 

 for the propeller. An error is admitted of at least 2 per 

 cent, of the maximum thrust in the calibration of the 

 dynamometer. The rotational speed of the propeller was 



I by means of a direct reading tachometer, and is 

 probably correct to about 1 per cent. Hut the power 

 absorbed was measured by assuming that the brake h.p. of 

 the Gnome engine, at a given speed, did not vary during 

 the course of a series of experiments. By this method of 

 measuring, we should estimate the probable error on the 

 measurement of power to be anything up to 10 per rent. 

 The speed of translation of the machine was mi '■■ 



means of an ordinary V tube, measuring the air pressure 

 in a converging rone. This was calibrated by Hying round 

 a measured aerodrome, taking the speed with a watch. So 

 that, taking into account the difficulty of flying exacts 

 over the course and of reading a water-gauge on a 



vibrating aeron the translational speed is probably not 



correct to closer than 3 per cent. 



niriitaVs stir les helices propulsives Aenenrrs.' Ry M. 



tagrand ( Bulletin tie la Societe d Encouragement pour l'lndustrie National-, 



\,„, I).;" 



