]\gt ; Energy Transformations of -Rays .
By Prof. W. H. BRAG , M.A. , F.RS .
, and H. L. PORTER , B.Sc. , Demonstrator of Physics , University of Leeds .
( Received May 15 , \mdash ; Read May 25 , 1911 .
) The manner of our attack upon the problems of radioactivity , including the action of the Rontgen rays , depends materially on whether suppose that an atom can or can not be made to yield energy from an internal store .
In the former case we have nothing to guide us to an estimate of how much energy may be expected to be put into circulation in this way .
If , for example , a -ray in passing through an atom prompts the atom to emit new secondary radiation with energy drawn from a source usually beyond the reach of transformation into physical or chemical or other known forms , or if the atom sometimes absorbs energy from that of the -ray motion and locks it up , then the quantities of energy thus added to or subtracted from the amounts we may hope to measure must be no more than subjects of experiment .
We shall have to be content with registering them without accounting for them .
But if we take the second of the two alternatives the problem is immensely simplified , and we may work for a more complete solution .
An -ray or -ray begins its career with so much energ ) .
While we cannot , of course , explain this initial liberation of energy , we may try to account completely for its subsequent expenditure in various ways , since we have no unknown or unexpected items to take into account on either side of the balance-sheet .
Exactly the same statement can be made in respect to each X- or -ray , since each such ray , as has been shown in previous papers , be considered by itself , being independent of all its companions in what we call a " " beam of X-or of -rays.\ldquo ; Energy considerations lead us directly to the supposition that the X- and -rays are corpuscular in nature in so far as each ray is a separate identity moving space unaltered in form and content , just as an unhindered projectile would do .
No X- or -ray spends energy in its passage through matter ; the only way in which the existence of such rays is made manifest is through their replacement by swiftly moving electrons which ionise the gas through which they pass .
The single X-ray disappears as such , and in its place is a cathode ray , an electron moving with energy inherited from the -ray .
Ionisation by X- or -rays is an indirect process .
* For a summary of the argument , see a paper by Brag , 'Phil .
September , 1910 .
Prof. W. H. Brag and Mr. H. L. Porter .
[ May 15 , It is not necessarily an exception to this rule that sometimes the absorption of a primary stream of -rays of definite quality or penetrating power is marked by the appearance of secondary X-rays of different quality .
Barkla has shown that many such cases exist , that the secondary rays are characteristic of the substance from which they come , and not of the rays which excite them , and that they can only be excited by rays more penetrating ( as measured by absorption in aluminium screens ) than themselves .
The corpuscular theory leads us to suppose that the more penetrating -rays as defined in this way are simply those with the most energy .
The quality of any -ray is completely defined by the energy of the cathode particle to which it gives rise .
At any rate there seems no ground at present for supposing any other qualification to be required .
Barkla 's law , then , amounts to this , that many substances under the stimulus of primary -rays emit secondary -rays which are characteristic of the substance and not of the stimulating rays , and that a secondary -ray can only be brought into existence by one having more energy than itself .
On the view of the energy question which we are adopting the last assertion in this statement is to be anticipated .
As we have said , this production of secondary -rays is not necessarily an exception to the rule that absorption of -rays implies the conversion of their energy into , that of cathode rays , and into nothing else .
It may well be that the secondary rays are due to a reconversion of the cathode rays into X-ray form , after they have lost energy in moving through the gas in which they arise , or in some other way .
Let us , for example , consider the history of a secondary -ray emitted by zinc under the stimulus of -rays from a Rontgen bulb .
It is an identity of definite energy , and is liable to be replaced in traversing some atom by a cathode ray of definite initial energy ; we may speak of the " " zinc X-ray\ldquo ; and the ' zinc cathode ray The electron of the zinc cathode ray is not different to other electrons , but it has a definite initial speed and a definite penetrating power .
Let us imagine a -ray penetrating a nickel plate ; in some cases the -ray is replaced by a Ni -ray , which is an identity of less energy .
This change may have taken place in one atom of nickel , which has then absorbed energy equal to the difference between the energies of a -ray and a Ni -ray .
Or there may be no exception*to the rule that -rays are replaced by cathode * In a recent paper Mr. J. Crosby Chapman ( Phil. Mag April , -1911 ) describes experiments which convince him that this cannot possibly be the process by which the secondary ionisation is excited .
Allowing X-rays to fall on bromine vapour he finds that the preseltce of or makes little difference in the amount of secondary radiation arising in the bromine ; whereas , since the must get in the way of many 1911 .
] Enerqy Transformations of -Rays .
rays only , and in that case we must suppose an intermediate stage , in which the -ray has been replaced by a moving electron ; in this stage some of the energy is frittered away in ionisation , and only a remainder handed on to the Ni X-ray .
We are here supposing that the X-rays are really homogeneous , and consist of a number of " " corpuscles\ldquo ; all alike .
The homogeneity is taken to be proved by the fact that they are absorbed , according to Barkla , under an exponential law , which naturally implies that , however many rays have been removed from a stream , the remainder possess the same properties .
Perhaps there is room to suppose that the quality of -rays is only an ayerage quality , but it is certainly difficult to understand the exponential absorption that case .
The foregoing statements are intended to be introductory to the description of certain experimental results .
We have attempted to consider and to measure the transformation of X-ray energy in a few cases , since the corpuscular theory would suggest the possibility of success in this direction .
If the absorption of -rays means their conversion into cathode rays , then we ought to find that the production of cathode ray energy is proportional to the disappearance of -ray energy , at least where there is no production of seco1ldary -rays .
In the latter case we may avoid the uncertainty as to the exact way in which these secondary rays arise , by for the energy contained in them .
Whatever energy of the primary rays is then left unaccounted for must be proportional to the cathode ray energy they can produce .
Lastly , we can show that -rays , when they go through a gas , produce enough cathode radiation to cause all the ionisation which is observed , and the1e is no need to ascribe any direct action to the -rays themselves .
We take first a pencil of the secondary X-rays from tin , which are relatively energetic , and give rise to penetrating cathode rays .
On this account the cathode rays which emerge from a plate on which the Sn cathode rays and the cannot do so , the CO : ought to interfere with the conversion of cathode rays into -rays .
This does not seem to me concIusive .
If the production of X-rays is a consequence of the encounters of cathode rays with bromine atoms there will be opportunity for the effect to take place even though the molecules are scattered among the Br atoms .
If an eIectron meets a molecule first it is not arrested there but deflected , and may have hundreds of encounters still before it : so that its chauce of meeting a Br atom is practically as great as ever .
If it is argued that the cathode ray is " " absorbed\ldquo ; by the bromine and the in proportion to weight , it must be answered that whatever " " absorption \ldquo ; may mean there is no clear evidence of the universality of Lenard 's law .
I do not assert that the idea of the intermediate corpuscular stage is established.in the simple form in which I have stated it ; but it se to simplify the theory , and is not , I think , disproved by Mr. Chapman 's experiment.\mdash ; W. H. BRAGa .
Prof. W. H. Brag and Mr. H. L. Porter .
[ May 15 , X-rays fall have a considerable ionising effect , for even those which have their origin in relatively deep layers of the metal can make their way into the open .
Our first object is to find the cathode radiation which the Sn -ray can arouse in different substances , and to compare the various amounts with the absorption coefficients of the substances for those -rays .
The apparatus is shown in fig. 1 .
A primary pencil of rays from the bulb passes down a tube lined with tin , and falls upon a plate of tin in a box made of tin .
The primary rays fall on nothing but tin , the object being to make the pencil of rays issuing from the box as free as possible from all but Sn -rays .
The secondary rays pass into an ionisation chamber through an FIG. 1 .
opening AA closed with paper .
They excite emergence cathode rays in sheets of metal laid on the paper , or incidence cathode rays in plates laid on top of the chamber if such sheets or plates are placed in position .
The ionisation current is balanced against a similar current in a second ionisation chamber crossed by a pencil of rays from the same radiator .
The second chamber is closed by an adjustable opening worked by a micrometer screw ; the values of the readings are determined by a special calibration .
A Wilson tilted electroscope is used as a detector .
The arrangement works well ; consecutive readings rarely differ by as much as lper cent. The results of experiment differ by far more than this at different 1911 .
] Energy of -Rays .
times , but it seems as if the bulb at fault rather than the measuring apparatus .
There are various ways of finding out what fraction of the ionisation current observed is due to the action of the cathode rays at incidence or emergence .
It may be done by gradually increasing the depth of the chamber , a method employed by Townsend* long ago .
Sadler .
has used the same method .
varied the air pressure , which is more convenient in ways , because it no moviDg parts in the apparatus and makes no change in the geometrical conditions .
In some cases it is convenient to increase gradually the thickness of the metal sheet , and to observe the consequent increase in the cathode radiation .
and MadsenS employed this method when investigating the production of -rays from -rays .
It is also very convenient to find the difference between the current when the metal under investigation is bare to the ionisation chamber , and the current when the metal is covered over with a very thin sheet of tissue paper , which cuts off all cathode rays and interferes very little with any others .
For this purpose we generally employ carefully flattened plates of metal , one face of each plate being bare and the other covered with tissue paper fixed on with as little adhesive as possible .
In order that results obtained at different times , or with different pieces of apparatus , may be comparable with each other , it is convenient to express each result as the ratio of the ionisation caused by the cathode radiation to the ionisation caused by the -rays in crossing 1 cm .
of air .
For example , in one experiment where the plate was made of Sn , with the bare side towards the chamber the current was 247 on an arbitrary scale , and when the plate was turned over so that the tissue paper cut off the cathode rays , the current was 122 .
Hence the cathode radiation was 125 , and , since the chamber was cm .
across , the ratio required was .
There was no secondary radiation in this case ; it is not excited by tin in tin .
In the same way we may find the corresponding figure for the " " emergence\ldquo ; cathode radiation .
In the case of Ni , Fe , Cu , and , the results are complicated by the presence of the characteristic secondary radiations .
The tissue paper no longer cuts off all the radiation from the metal ; the cathode rays are absorbed , but the secondary -rays should be very little affected in passing through the paper .
To make sure that this was the true explanation we PhiL Soc. Proc 1899 , vol. 10 , p. 217 .
'Phil .
Mag March , 1910 . .
Phil. Soc. Proc , vol. 15 , p. 416 .
S PhiL Mag December , 1908 .
Prof. W. H. Brag and Mr. H. L. Porter .
[ May 15 , stretched a piece of paper over the top of the ionisation chamber so as form the upper wall of it .
The ionisation current then observed was nol increased when plates of tin or aluminium were placed over the chamber on top of the paper , but it was increased when any one of the other metals was placed there .
By interposing thin sheets of aluminium or paper between the plate and the paper top to the chamber , it was possible to find an absorption curve .
This curve is not exponential ; it is of a more complicated form , which we must examine presently when we compare the energy of the secondary radiation with that of the primary .
At this stage it is only necessary to say that when the plate on top was a copper one the curve showed an absorption such as would be expected if the secondary -rays were Cu -rays .
We may assume , therefore , that we have here an example of the effect described by Barkla , and that the secondary rays excited by the Sn rays are characteristic of the metal in which they arise .
The amount of the secondary radiation is conveniently expressed by the ratio of the ionisation it causes in the chamber to the ionisation caused by the primary rays in the same chamber .
For example , in an experiment when the rays entered the chamber through a paper screen and fell upon an aluminium plate with the papered face down , the current was 136 .
When nickel with the papered face down was substituted for the aluminium , the current was increased to 177 , and if the nickel plate was turned er so as to present its bare side to the -rays the current became 296 .
Thus there was an incidence cathode radiation of 119 , and a secondary -ray effect of 41 .
The former is recorded as and the latter as 41/ 137 or The ratio of the emergence to the incidence cathode radiation was determined separately .
The following table ( p. 355 ) sho ws a set of results obtained in this way .
In the case of zinc the ratio of e1nergence to incidence cathode rays was not actually measured , but assumed to be the same as for copper .
The total cathode ray effects and the absorption coefficients ] run parallel to each other , allowing for the fact that in the case of some of the metals part of the primary -ray energy is spent in exciting secondary X-rays .
But there are several points to be considered before a true rison can be made .
In the first place the cathode ray effects are caused by such of the cathode radiation as emerges from a plate on which the rays , fall .
This is not exactly what we want ; we are trying to compare the cathode radiations excited in equal weights of different metals by Sn -rays , and we have really no measures of these as yet , for we have not allowed for the Energy Transformations of -Rays .
Table I.\mdash ; Sn Rays .
Column I refers to the metal in which the X.rays excite cathode rays .
II gives the intiidence cathode ladiation expressed as already explained .
III gives the ratio of emergence to incidence cathode IV gives the total cathode radiation .
V gives the secondary -radiation expressed as already explained .
VI gives the absorption coefflcient of the Sn rays in each metal .
loss of energy of the cathode rays in their way out of the body of the metals .
If the proportionate loss is the same for all metals , it is unnecessary to find it , and there are rounds for believing that in many cases it is .
Lenard showed that the cathode rays which he employed were equally absorbed by all substances with atomic weights between those of carbon and gold .
The value of the coefficient which he found was 3000 nearly , so that his rays must have been about as penetrating as the Sn cathode rays .
* Hydrogen was an exception , having a larger coefficient .
Lenard also compared the absorption coefficient of less penetrating cathode rays in some gases , of which the only one with a large atomic weight was argon .
Though the coefficient was still much the same in , and , the atom had increased its difference .
These experiments do not cover the penetration of the slower cathode rays in metals such as Fe , Cu or ; and we do not know that the law of equal penetration holds ill these cases .
They are very difficult to investigate since the absorption coefficients are so high .
Confining , however , our attention to the Sn rays for the time being , it seems right to assume Lenard 's law to hold to a first approximation .
If we do this , the amount of cathode radiation emerging may be taken as a true relative measure of the amount of cathode radiation made .
We must now try to allow for the energy spent in making secondary X-rays .
Barkla has shown that suoh secondary rays are radiated uniformly in all directions about their origin .
Let a pencil of rays of energy I enter the upper plate of the ionisation chamber and let be the absorption coefficient of the plate for these rays .
Let be the absorption coefficient *Beatty , .
cit. Prof. W. H. Brag and Mr. H. L. Porter .
[ May 15 , for the same plate in to its own rays .
The loss of energy of the primary in crossing a of weight is , where is the mass penetrated previously .
Of this the amount that gets back into the chamber is where is the fraction of the primary energy which is directly or indirectly converted into secondary X-ray energy , and the factor in brackets is necessary to allow for the absorption in the stratum of weight when the rays are distributed uniformly , having been obtained by measurement with rays normal to the absorbing plate .
* The divisor 2 is introduced because half the energy must go each way .
The whole of secondary radiation which comes back and crosses the chamber is therefore and is a function , therefore , of The value of the integral can be found for a few values of by plotting the curve and finding the area .
The result is shown in fig. 2 , and 'Phil .
Mag. , 'May , 1910 , p. 735 .
Values of the constant have been already found for a number of cases by Sadler , ' PhiL Mag July , 1909 .
1911 .
] Energy of -Rays .
a smooth line drawn through the calculated points gives us the value of the integral sufficiently well for all values of In the case for example of Sn rays incident on Cu , since and so that , the value of the integral is found from the curve to be ; or Now the absorption coefficient of Sn rays in is nearly , and of Cu rays in is , and the ratio of the coefficients in air will be the same practically as in C. Also the ionisation of the air is proportional to the absorption coefficient of the air ; hence By actual experiment the ratio of these ionisations is found to be and therefore .
That is to say , of the absorbed energy of the primary is spent in producing secondary Cu rays , and only of the energy remains to be accounted for in cathode ray } .
There is still one correction to be made , which is only of importance , however , in the case of aluminium .
When we measure cathode rays in this way we measule actually the difference between the cathode radiation from the etal in question and the cathode radiation from the tissue paper with which we have covered one face of it .
The results we obtain are therefore comparable with the difference between the absorption coefficients of the metal and the paper .
In the case of Sn rays the absorption coefficient of [ carbon is about .
Making all these allowances we obtain a set of results which are set out in Table II .
On the whole the figures in the last column agree very well , quite as closely as we can expect considering the sources of uncertainty which we have pointed out .
So far , therefore , we may conclude that the relationship between the cathode radiation and the absorption coefficient is as it should be .
Experiments with arsenic X-rays gave the results set out below in Table III .
Only the incidence cathode rays were taken into account , for the measurement of emergence cathode rays is unsatisfactory unless the plate through which the X-rays enter the chamber is very thin indeed .
The thinnest sheets obtainable of most substances were sufficient to harden the X-rays a little in spite of their approximate homogeneity , and there was a certain small fraction of scattered primary X-radiation included in the secondary radiation from the arsenic .
A very little penetrating radiation in VOL LXXXV.\mdash ; A. 2 Prof. W. H. Bragg and Mr. H. L. Porter .
[ May Table II.\mdash ; Sn Rays .
Column I. Name of metal in which the cathode and secondary -rays are excited .
II .
Amount of cathode radiation measured as described above .
III .
Amount of secondary -radiation .
IV .
Calculated value of ratio of secondary to -ray energy .
V. Absorption coefficient of primary X rays by the metal .
VI .
PropQrtion of the absorption which should go to the production of cathode VII .
The same , less the absorption coefficient of paper ( see above ) .
VIII .
Ratio of VII to II .
a fairly homogeneous beam produces a material alteration in the cathode radiation .
For example , Beatty 's figures on p. 324 of the paper already quoted and published in the ' Phil. Mag August , 1910 , show that a stream of Sn -rays producing the same efiect in air as a stream of As -rays produces seven times as much cathode radiation in silver .
As it happens , there is very little difference between the emergence and incidence cathode rays in such ctises as we have examined , and it seems justifiable therefore to work with the incidence rays only .
The arsenic radiator was made of the same form as the tin radiator ; tube , box , plate and all parts on which the primary rays could fall were made of thick card on which metallic arsenic in a powdered form had been fastened by a little gum .
Perhaps the gum scattered a little hard prinlary radiation , for our absorption coefficients were somewhat smaller than those of Barkla and Sadler .
We know hardly anything of the absorption coefficients of As cathode rays .
Assuming that they also follow Lenard 's ] , the figures in the last column should be all the same .
They are so approximately .
Some results obtained with Zn rays may be given .
As with arsenic rays , they are difficult to interpret fully until more is known of the absorption of the special cathode rays in different metals .
Ihere is also more liability to experimental error since the cathode radiations are so small in many cases .
We have , therefore , following Beatty 's plan , worked at low pressures so as to make the cathode ray effect , reater relative to the effect of the -rays .
Even then it is difficult to find accurate results , for when zinc rays fall on a metal such as aluminium Prof. W. H. Brag and Mr. H. L. Porter .
[ May 15 , covered with tissue paper and the broken one shows the effect when the metal is bare .
The lines are drawn straight though clearly they should all be slightly curved in order to fit the points ; the figure seems easier to examine with the lines drawn as they are .
The results are collected in the following table:\mdash ; Table Rays .
For description , see notes to Table II .
The parallelism between the absorption coefficients and the production of secondary radiations is quite clear ; and when due allowance is made for the energy which reappears as secondary -rays , there is a fair agreement between thep production of cathode rays and the energy which presumably goes to make them .
Copper does not agree with the others ; and we have no reason to as yet for the divergence .
It is to be remembered that agreement cannot be expected unless cathode rays are equally absorbed in all substances , and we no sure knowledge that they are .
Beatty*has traversed the field in a cross direction .
He has allowed -rays of various types , viz. , those from Fe , Cu , , As , and Sn to pass through a silver leaf , and has measured the " " emergence\ldquo ; cathode radiation arising from the silver in each case , expressing the result as a ratio of the ionisation caused by it to the ionisation caused by the -rays in a layer of air 1 cm .
thick immediately after through the silver .
It is necessary , of course , to use some such standard of reference .
The ratio is called in Table I of his paper ( p. 324 ) .
He also measured the absorption coefficient ( ) of the cathode rays in air by finding the thickness of air stratum required to reduce the cathode radiation to half value ( in the same table ) and dividing this into .
This process depends on the assumption of an experimental law of absorption .
The absorption does not really fulfil this law , as his own diagrams show ; but for the purpose of argument it was no doubt justifiable to make 'Phil .
Mag August , 1910 .
1911 .
] Transformations of ] the assumption for the sake of convenient expression , and he was concerned only with the relative amounts of the cathode radiations produced in silver by the different rays .
If the assumption of the exponential law made an error in one case , it probably made the same error in the others .
He then assumes , further , that the in silver is proportional to the in air for all these types of X-rays , so that while is a measure of the emergence cathode radiation only , is a measure of all the cathode radiation that is produced , not of what gets out of the silver .
It is the ratio of the ionisation due to this radiation ( if it could all get out ) the ionisation due to the -rays in 1 cm .
of air .
He assumes , still further , that the ionisation of the air by the -rays is proportional to the absorption of those rays by the air ; and , again , that the absorption of -rays by the air is proportional to their absorption by silver .
The constancy of the quantity therefore enables him to draw his final conclusion that , " " in the case investigated , a constant fraction of the homogeneous radiation is spent in producing cathode particles Some of Beatty 's assumptions are certainly justifiable , the third , at auy rate , and probably fourth ; the first is permissible for the occasion , and the second may very well be true , but there is no definite proof of it .
The corpuscular theory would lead us to anticipate Beatty 's result : it would , indeed , further still , and assert that the constant fraction is unity , at least in the case where there is no secondary -radiation in addition to the cathode radiation .
There was none in Beatty 's experiment .
If we aim at an experimental verification of this simpler but more comprehensive deduction from the corpuscular theory we cannot assume an exponential law for the absorption of the cathode rays in the silver .
We must make some attempt to find the true law .
We pass Sn rays into an ionisation chamber through an opening closed by a card .
The whole of the chamber , which is made of brass , is lined with Al and paper , so that there is no radiation from the walls to be taken into account .
Silver foils then laid on the card , and the increase in ionisation is observed as the number of foils is increased .
The curve obtained is shown in .
If we can find the slope of this curve at the origin we shall know the cathode radiation for extremely thin sheets , which is what we really require ; for if the sheets were very thin all the cathode rays would get out .
Unfortunately silver foil is not to be had thin enough to any points nearer the origin than are shown in the figure , and we are left to draw the curve through the experimental points as we think it ought to be drawn .
Beatty assumed the exponential curve passing through the point for which the current measured is half the final value .
This is represented by the Prof W. H. Brag and Mr. H. L. Porter .
[ May 15 , dotted line in the figure , and OQ is the tangent at the origin .
The true curve would seem to be somewhat different and less steep initially .
The simplest form to assume is parabolic , viz. , , and such a curve has been drawn so as to pass through the origin and the first two points ; it then FIG. 4 .
breaks away as shown .
Adopting this form , it is easy to find the at the origin , OP ; and it appears that initially emergence cathode rays are produced and make their way into the open at the rate of 208 to a weight ( of screen of .
The maximum reading due to cathode rays is 309 .
We derive from this curve the fact that when the maximum emergence radiation is 309 , the actual production of emergence cathode rays is or 832 per milligramme of silver .
The ratio of these two numbers , , and is all that is required for future calculation .
If we had assumed an exponential curve passing through the point on a curve at which the ordinate is half its greatest height we find that the tangent at the origin , , gives a value for this ratio about 15 per cent. higher .
In order to find the full amount of the cathode radiation produced in silver we must also take into account that which would be found on the incidence side of a very thin sheet .
We must do the same thing in regard to incidence cathode rays as we have just done in regard to emergence .
This might be done by gradually thickening a silver sheet on top of the chamber on which the X-rays fall : but there is always some uncertainty about this method , since some of the -rays which enter the chamber might \mdash ; through scattering or otherwise\mdash ; strike the walls than the top plate .
When dealing with the emergence cathode radiation no such thing can happen .
It is perhaps simpler to compare the incidence and emergence cathode rays by laying one or more silver foils on a card and placing this in 1911 .
] Energy Transformations of .
363 the middle of the ionisation chamber , so that the -rays must go through it .
The cathode radiations are then found by subtracting the current for a bare card from the current when there are silver foils on one side or the other .
We found the ratio between incidence and emergence to be for one , two , or three foils ; and we conclude that this ratio may be supposed to hold for extremely thin films .
The distribution of the cathode rays about the silver atom is therefore unsymmetrical from the first , though it is by no means so uni-directional as in the case of the -rays caused by -rays .
The measurements made so far were obtaiued in air , because it was con- venient to adjust the silver foils in a chamber which could be easily opened and closed .
But the remaining part of the experiment was carried out with oxygen instead of air , since it was necessary to compare the absorption coefficient of the Sn rays in the silver and in the gas employed .
It is difficult to measure accurately the absorption of Sn rays in a gas since it is so very small , and air can only be used as gas .
But the absorption in oxygen can be found by using solid absorbing screens containing known quantities of oxygen .
This part of the experiment will be described presently .
It was necessary to find the ratio of the emergence radiation from the silver , measured by the ionisation it causes in oxygen to the ionisation in a given layer of oxygen due to the Sn -rays .
Employing the same method as Beatty , we passed Sn rays into an ionisation challlber which could be filled with gas at any pressure .
The rays passed in through ten silver foils , which were amply sufficient to give the full cathode radiatiolL The depth of the chamber was cm .
The curve obtained is given in fig. 5 , and gives us Prof W. H. Brag and Mr. H. L. Porter .
[ May 15 , the means of calculating the ratio we require .
The ionisation due to the emergence cathode rays is 344 , while that due to the -rays crossing cm .
of oxygen at 76 cm .
pressure .
Hence the ratio of the effect of the emergence cathode radiation to the effect of the -rays crossing 1 milligramme of oxygen is the density of oxygen being We have already found that the cathode radiation per milligramme of silver is times the emergence cathode radiation from silver , result which was found by the use of air , but must be just the same when oxygen is used .
Combining these figures with the ratio of the whole cathode radiation to the emergence radiation ( viz. , we have finally that in equal weights of silver and oxygen there is times as much* ionising power in the cathode rays produced by the X-rays in the silver as there is ionisation actually effected by the -rays going through the oxygen .
We nexb require the ratio of the absorptions of the Sn rays in silver and in oxygen .
Using the same apparatus and keeping all the conditions the same in the absorption experiment as in those that went before , we found the absorption coefficient of the -rays by the silver to be .
The very soft portion ' of the Sn rays was intercepted by a wooden plate which closed the ionisation chamber , and was used throughout because it stood air pressure without cutting out many Sn rays .
The absorption coefficient of was found to be and of oxalic acid to be .
From Beatty 's results the absorption coefficient of is negligible .
Hence we can calculate the absorption coefficient of oxygen from the equation and therefore Consequently silver absorbs Sn rays , or times as much as oxygen .
From the earlier results in this paper we therefore conclude that Sn rays produce times as many thode rays in silver as in , so that the ionising power of the cathode rays produced by the Sn -rays in 1 milligramme of silver is times the ionising power of the cathode rays produced by the same -rays in 1 milligramme of oxygen .
Comparing this with the result of the last paragraph we see that the ionisation effected in * That is to say , the total ionisation they would cause in oxygen if they could complete their whole courses in the gas .
For , allowing for water of crystallization , there is one carbon atom to three of gen and three of hydrogen .

