Oct. 3, 1878] 



NATURE 



605 



Experiment III. — Arrange, as in Experiment I., but connect 

 with a gas supply and ignite. The primary expansion will be 

 obtained but not succeeded by any other. This is readily 

 accounted for. The edges of the expansion, as far as the middle, 

 are receding from each other, and there being no cohesion in 

 gases, they continue to recede, and so the expansion is fan- 

 shaped. 



Ex/en'ment IF.— Sta.rtirig with the jets at an angle of about 

 60° and their ends nearly in contact, lessen the angle and adjust 

 the gas supply, that the flame may take the shape shown in 

 Fig. 3. At a certain point increase of gas will cause the flame 

 to emit a note (usually a very high one), and at the same time 

 commencing somewhere about o^ there will be an expansion at 

 right-angles to the primary one. A 

 horizontal section of the flame will now 

 have a cruciform contour. 



Having obtained this musical flame, 

 withdraw the jets through their corks 

 (keeping the same angle throughout). As 

 the distance between them increases the 

 note becomes of lower pitch. With careful 

 attention a very low note may be obtained. 

 At a certain distance the note ceases, but 

 may be restored by additional gas pres- 

 sure. 



JVo^e. — It is evident that the velocity of 

 the gas decreases with its distance from 

 the jet. Hence the greater the velocity 

 the higher the note. 



I have shown (Nature, vol. xv. p. 

 119) that the velocity of the escaping gas 

 is decreased by ignition, therefore if a 

 certain flame produces a low note the 

 same column unignited will give a higher 

 one. This is found to be the case. 



Experiment V. — Having arranged, as 

 in the preceding experiment, reduce the gas supply till the flame 

 ceases to sing. It will now be sensitive to a certain note. 

 Produce this note very near the flame, the flame takes the same 

 shape as when singing (Experiment IV.). 



Experiment VI. — Produce a flame tolerably "sensitive (Expe- 

 riment V. ). Mount a whistle so that it may revolve round the 

 flame. In two positions, being the ends of a diameter passing 

 through the edges of the flame, the sound will produce no effect. 

 Note. — It will be seen that the sonorous impulses strike the 

 flame in such a manner that the two columns are thrown into the 

 same phase of vibration. The maximum effect is produced at 

 points at right-angles to the minima. The waves reaching 

 the nearer column first will throw it into a phase dif- 

 fering from the distant one. 



Experiment VII. — Select jets of such a size that a 

 flame may be obtained as large as one's hand. At a 

 pressure just below that at which it sings it will be 

 highly resonant.^ 



A soimd produced in its vicinity will cause it to 

 resound. When the exciting note has ceased, the 

 sound of the flame will gradually subside ; at the same 

 time giving forth higher and higher harmonics of the 

 note. 



The sounds, however, are very feeble, and can only 

 be heard at a very short distance. A convenient 

 arrangement for amplifying the soimd, is to place the 

 stoppered end of a large gas jar in contact with the ear, 

 and direct the open mouth towards the flame. With 

 a correct adjustment any note within a very wide range 

 will excite the flame, which may emit the same, or 

 some harmonic of it. By whistling an air with the 

 mouth, a rather pleasing accompaniment is heard, and 

 the extreme gravity of som.e of the flame-notes is certainly 

 remarkable. I have not succeeded in a;igmenting the notes so 

 as to make this a lecture experiment. 



Fig. 4 consists of a cork or other suitable material about 6 in. 

 long, and i in. wide and thick. The axis of a jet A is directed 

 to just pass clear of the top of a jet B. A is supported in a cork 

 so that it may approach or recede from B. The end of A is 

 contracted till it will give a flame about i \ inch long, imder the 

 full pressure of the gas mains (yV in. water). 



' This corresponds with the intersection of axes of the jets. 



' A large gas supply is indispensable, and very careful adjustment is 

 required to obtain the^most favourable result. A pressure of ij inches of 

 ■water is sufficient. 



When tested with water it should show very slight segmenta- 

 tion. 



Experiment VIII. — Produce at B the smallest possible flame, 

 and direct the full flame from A, so that the point just passes over 

 the top of B. Extinguish A without turning off the gas. The 

 issuing gas will form a bluish cone beyond B, the space between 

 A and B remaining unignited. 



A may now be withdrawn from B till the cone becomes un- 

 steady. If the column has very slight segmentation the distance 

 may be five or six inches. With most jets, however, the limit 

 will be about three or four inches. 



This column is exceedingly sensitive. The faintest sound 

 to which it responds will cause the ignited cone to recede 

 towards A. 



The greater length, and therefore the greater velocity of the 

 unignited column will be at once noticed [pide Experiment IV.) 



Experiment /A!".— Arrange as Experiment VIII., and 

 remove A just so far, that the cone does not strike back to it. 

 Now very softly produce the responding note. The column 

 recedes and becomes ignited through the whole of its length. 



Experiment X. — Replace A by a rather more segmental jet, 

 and obtain the cone. It will be seen that when the cone 

 recedes, it is divided at the extremity more or less perfectly, into 

 two parts. 



Experiment XL — Replace the jet B by a disc of spongy plati- 

 num ^ about the size of a halfpenny. 



The unignited column causes an annular patch of the platinum 

 to become red hot. When the column is excited the annulus is 

 divided into two spots. The same column, if previously ignited, 

 will not be affected by the same sound. 



This forms another proof of the excitability of the unignited 

 column. 



In Fig. 5 two equal jets inclined to each other at an 

 acute angle, are fixed in a cork which freely slides in a tube 

 S in. X 4 in. 



Experiment XII. — Using the apparatus (Fig. 5), slide the 

 cork near the top of the tube and ignite the gas. Slowly lower 

 the cork till a point is reached at which the flame is sensitive. 

 Observe that the base of the quiescent flame in contact with the 

 tube forms a sinuous line, consisting of two depressions and two 

 crests at right angles to each other ; and further, that the crests 

 correspond in position with the edges of the primary expansion, • 

 and the depressions consequently with the sides. 



Note. — When the flame is excited their relative positions are 

 reversed. 



Experiment XIII. — Use a low gas pressure (one or two 

 tenths) with Fig. 5, and slide the cork down the tube till the 

 base of the flame becomes unsteady. At a certain point a 



noise will cause the flame to move from side to side in the 

 same plane as the jets. Increase of pressure accelerates the 

 movement. 



Instruments for the production of sensitive flames may be 

 divided into two classes : — 



1. Those in which the whole of the gaseous column is ignited. 



2. Those in which the column is only partially ignited. 



' A good substitute for the spongy platinum is obtained by pounding some 

 fine asbestos till quite smooth, moistening with a tolerably strong solution 

 of platinum tetrachloride, moulding the mass into the required shape in a 

 piece of paper, and ig^niting. The paper bums off and leaves a porous 

 fragile mass. This is 5temately soaked in the platinum solution and ignited 

 three or four times, when it becomes tolerably durable. 



