]\gt ; ROYAL SOCIETY , MARCH 19 , 1914 .
Discussion on the Structure of the Atom .
Abstract of Opening Address by Sir E. RUTHERFORD .
When the Secretary of the Royal Society asked me to open a discussion on the structure of the atom , I very gladly acceded , for I felt that it was a snbject which would be of interest to many members of the Society , even if their own investigations are not directly connected with this special field of enquiry .
I should first like to point out that a large part of the work in modern Physics for the last 20 years has been intimately connected with the problem of the structure of the atom , and several very powerful methods of attack have ecently been developed to throw light on its structure .
The discovery of the negative electron as a constituent of all atoms much widened conceptions , while the detailed study of radioactive substances has brought to light the remarkable series of transformations that occur in them , resulting in the addition of more than a score of new " " transition\ldquo ; elements , which have a limited and break up with the emission of intense radiations and of a large amount of energy .
A of the scattering of swift and -particles in passing through matter has thrown much light on the structure of the atom , for these particles have such great energy of motion that they actually pass through the individual atoms , and are deflected by the intense electric fields they meet within the atom .
In addition , the recent methods of studying -ray spectra by their reflection from a crystal promise to much valuable information on the modes of vibration of the inner constituents of the atom .
The proof that the negative electron was a constituent of atoms , and could be released from them by a variety of agencies , gave a great impetus to the study of atomic .
It was soon recognised that a heavy atom must be regarded as a complex structure , consisting of negatively and positively charged particles , which held in equilibrium mainly by electrical forces .
A type of atom model , suggested by Lord Kelvin , has been developed and worked out with great detail and skill by Sir J. J. Thomson , and is now generally known as the " " Thomson\ldquo ; atom .
The atom is supposed to consist of a sphere of uniform positive electrification extending over a distance comparable with the ordinary accepted diameter of the atom , and containing within it a sufficient number of electrons to make it electrically neutral .
This type of model atom has the great advantage that the distribution and velocities of the component electrons which make up the atom can be worked out mathematically in detail , and it has been shown that the eIectrons tend to distribute themselves either in rings in one plane or in spherical shells like the coat of an onion .
Sir J. J. Thomson has shown that such a model atom imitates in a way several of the general properties of the actual atoms , and affords a suggestive explanation of the reason why some atoms are electropositive and others electronegative , and of the variation of the properties of atoms with atomic weight as exemplified in the Periodic Law .
( 8001 ) There is one difficulty , however , that is common to all types of model atoms .
According the electromagnctic theory , electrons in orbital motion must radiate , but this difficulty has to some extent been removed by the proof of Sir .
J. Thomson that the orbital motion of a large number of electrons in a ring radiates eItergy at a very slow rate .
Although the atom may remain stable long periods , it must uItimately break up , due to this uous drain of energy by radiation .
While the " " model atom " " of Sir J. J. Thomson has performsd a very useful service in indicating the general erties of the electrical atom , and has proved very amenable to calculacion , it does not seem able without modification to explain some experimental facts that have been brought to light in recent years .
It has now been definitely established that when swift -particles pass through a thin film of matter , some of the -particles are deflected through more than a right angle in consequence of the close encounter of an -particle with a single atom .
This result shows that the atom must be the seat of intense electric or magnetic forces .
General calcuIation shows that the particular distribution assumed the positive electricity in the Thomson atom does not admit of sufficiently strong fields to deflect an -particle through .
large angle , unless it be supposed that the positive electricity is concentrated in a very small volume .
In order to explain such results , I was led to consider the properties of a " " nucleus\ldquo ; atom , in which the positive charge and most of the mass of the atom are concentrated ithin an exceedingly small volume at the centre of the atom .
The nucleus is surrounded by a distribution of electrons to make it electrically neutral , but.extending for a distance from the centre comparable with the diametsr of the atom .
It was supposed that the large single deflection of -particles was due to their passage close to the nucleus , where the field is very intense , and where few , if any , electrons are present .
Simple calculation shows that the external electrons are incapable of producing more than small scattering of the -particle passing through tlJem .
Regarding the -particle and nucleus as point charges of electricity , the laws of this large-angle scattering were simply deduced on the assumption that the forces between them were electrical and followed the inverse square law .
These were examined in great detail in the experiments of Geiger and Marsden .
They found , for example , that the number of -particles scattered through different angles agreed with the theory over a very wide range , and in general they verified all the main edictions of the ory .
Darwin has shown that no other law of force , except the inverse square , is consistsnt with the experimental results .
The scattering of high-speed particles like -partioles , in their passage through the atom , thus provides us with a powerful method of deducing the intensity of the forces in the atom , and thus throwing light on the structure of the central part of the atom .
From the magnitude of the scattering at different angles , it is possible to deduce the nucleus charge .
The value of the was found to vary approximately as the atomic weight , and to be equal to about , where A is the atomic weight and the unit of .
On this view , the number of external electrons in the atom is about one half of the atomic weight in terms of hydrogen .
It is signfficant that a similar result was obtained by Barkla from consideration of the scattering of X-rays by matter .
These conclusions were based on the theory of scattering of X-rays by electrons worked out by Sir J. J. Thomson .
From these results , we may consequently conclude that the nucleus charge of gold , for example , atomic weight 197 , is about 100 units , and tlJat it is surrounded by 100 electrons .
( Note.\mdash ; The valus of the nucleus charge deduced from its X-ray spectra was shown by Moseley the discussion to be 79 .
) From the radioactive evidence , it appears probable that the -partide is the nucleus of helium and contains two positive charges , and consequently ogen can only have a nucleus charge of unity .
Van den Broek and Bohr have lggested that the nucleus charge may prove to be the atomic number of the elements wlIen arranged in increasing order of atomic weight .
On this view , which has much to commend it , hydrogen has one charge , helium two , lithium three , carbon six , and so on .
This important point will be referred to later .
From the scattering of the -particles by gold , it was deduced that the -particle must approach to within cm .
of the nucleus to be deflected through nearly .
This suggested that the nucleus of the atom has dimensions exceedingly minute red with the sphere of action of the atom , and that , consequently , the positive electricity is very concentrated .
This view is strongly Qupported by some recent experiments of Marsden .
On the theory , an -particle in a close encounber with a hydrogen atom should set the h.vdrogen atom in very swift motion .
Under the best conditions , the hydrogen atom should gain a velocity about times that of the colliding -particle .
To test this point , an intense beam of -particles was passed through hydrogen .
Most of the -particles were stopped at 20 cm.\mdash ; the range of the -particles in hydrogen , but a small number of scintillations on a zinc sulphide screen were observed for a distance up to 90 cm .
No such effect is observed in air .
There appears to be no doubt that some of the hydrogen atoms are set in such swift motion by the -particles that they have a range in hydrogen about four times greater than the -particle , and are able to produce visible scintillations .
-simple calculation shows that the -particle must approach within cm .
of the nucleus in order for the latter to gain its great velocity .
Now this distance is very small , even smaller than the accepted diameter of the electron , cm .
This calculation gives a maximum limit to the dimensions of the hydrogen nucleus , and it may be much smaller .
For these reasons , I have suggested that the mass of the hydrogen atom may be electrical in origin and that the nncleus is the positive electron .
We shall now consider the bearing of the radioactive evidence on the structure of the atom .
The expulsion of charged helium atoms in so many of the transformations indicates that the nucleus is composed , at any rate in part , of helium atoms .
The energy of expulsion of -particles can be readily explained by its passage through the strong nucleus field even if it escapes with little initial velocity .
The general evidence indicates that the primary -particles arise from a disturbance of the nucleus .
The latter must consequently be considered as a very complex structure on a minute scale consisting of positively charged particles and electrons , but it is premature ( and would serve no useful purpose ) to discuss at the present time the possible structure of the nucleus itseIf .
On the nncleus theory , the ordinary properties of the atom are dependent entirely on the nucleus charge , since the value of the latter must determine the electronic distribution .
The properties of gravitation and radioactivity which are unaffected by chemical or physical agencies must be ascribed to the nucleus .
( 8001 ) Recently Faians and Soddy have put forward a very important generalisa to explain the iation of chemical properties of the series of elements resulting from the transformation of the radioatoms .
They suppose that the expulsion of an a-particle changes the position of the resuIting element two groups in the periodic table in the direction of diminishing mass , while an expulsion of a -particle changes it one in the opposite direction .
On the nucleus theory , this supposes that the loss of an -particle ( carrying two ullit positive charges ) lowers the nucleus charge by two units , while the loss of a -particle ( carrying one negative charge ) raises the nucleus charge by one unit .
On this view , for example , the elements radium and radium should have the same nucleus charge as the end-product of the radium series , which is believed to be lead .
It is significant that both radium and radium are so closely identical in chemical properties with lead that they are inseparable from it .
Numerous examples of a similar kind are known amongst the radioelements .
We are thus led to the important conclusion that elements of different atomic weights may have identical chemical and physical properties .
If the nucleus charges are the same , they should also give nearly identical spectra , for it seems unlikely that the mass of the nucleus will exert much influence .
No doubt definite evidence on these points will be soon forthcoming .
A number of cases are known in the radioelements where the nucleus charge differs by unity .
Recently some very evidence on this point has been obtained by Moseley by entirely distinct methods for a number of ordinary elements .
As you know , experime1ltal methods have recently been developed by Brag , Moseley and Darwin for determining the -ray spectra of slements by reflection from crystals .
The frequency of vibratio1l , which is about 1000 times greater than for the shortest light waves , can be directly deduced from the structure of the crystal .
When an element is exposed as anticathode in an -ray tube , it emits its characteristic radiation and its spectrum is examined by the crystal method .
Moseley examined in this way the -ray spectra of twelve consecutive elements from calcium to nickel .
The spectra of all these elements are similar in type , consisting essentially of two strong lines .
The frequency of these lines was found to increase by definite steps in passing from one element to the next .
Without discussing the theory of the origin of these spectra , it suffices to say that the frequency was found to be proportionaI to where was a rohole number and a constant about unity .
in this case was a whole number which varied by unity from 20 for calcium to 30 for zinc , and corresponded to the atomic number or nucleus charge .
These results thus show clearly that the mode of vibration of the interior of the atom is closely connected with the number which is believed to be a measure of the nucleus charge .
I understand that Mr. Moseley has still further experimental results to contribute later in the discussion .
I have far not dealt with the important question of the distribution of the external electrons which make up the atom .
It is obvious thsre are great difficulties on the ordinary mechanics in considering possible distributions , for there is no reason why an electron if disturbed should not always fall into the * Since the discussion Dr. Andrade and I have that the strong lines of the -ray spectrum of radium are identical with the characteristic X-ray of lead .
13vidence has also been obtained by several investigators that the atomic weight of lead formed from radioactive matter differs from that of ordinary lead .
nucleus .
In order to ercome these difficulties , Bohr has supposed that the electron may take up a number of definite positions with regard to the nuclsus , corresl)onding to what he calls ' the stationary states A radiation of definite equency is supposed to be emitted when an electron falls from one stationary sbate to another .
By introducing a conception connected with Planck 's quantum , he has attempted to explain the of the series spectra of the simple elements , and has shown that the fundamental frequency is in close accord with experiment .
The atom of hydrogen is supposed to be the simplest possible structure consisting of a nuclens of unit charge and one electron .
The complexity of the spectrum in this case is supposed to be due not to the complexity of the atom but to the number of possible stationary states which the electron can take up .
In a similar way , he supposes the atom of helium to consist of a nucleus of two charges and two electrons , and has given reasons for believing that some recently discovered spectra which have been attributed to hydrogen in reality belong to helium .
This question is , however , still subjudice .
The structure of a number of elements has been indicated along these lines , supposing that the nucleus oharge increases by unity in passing from one atom to the next .
Bohr has also discussed the structure of the hydrogen molecule , which is supposed to consist of two hydrogen nuclei with a of two electrons rotating between them .
The heat of tion has been calculated and is in fair accord with the Iatest experimental data .
He finds that the helium atom as constructed refuses to unite with another helium atom with a hydrogen atom .
This is in agreement with the observed behaviour of helium .
While it is too early to say definitely whether the theories of Bohr are valid , his contributions to the structure of simple atoms and their spectra are of great importance and interest .
In this address I have e1ldeavoured to point out briefly the main of attack on this question , and the point of view that at present guides ths investigations of others in this most difficult and interesting of problems .
Mr. H. MOSELEY : The chief difficulty at the present moment in lssing the structure of the atom is that the experimental methods by which we can get anywhere near to the centre of the atom are extremely few .
We may say that Sir Ernest Rutherford 's view of the atom arose originally and is based principally upon the one all-important fact of the scattering of the -particles .
The -particle is the only which can penetrate sufficiently deeply inlio the atom to find out about the central part of the atom .
If we are ever going to find out what is the structure of the atom , it will be from the middle outwards .
I want to show you that the -ray spectra of the elements give a rather hopeful method of adding to our information .
The possibility of analysing the -ray spectra we owe to the work of Prof. Brag and his son .
The only way in which the work which I have followed out differs from theirs is that , while they used electrical methods , I have registered the spectra photographically .
It happens that when a very large number of elements have to be rapidly investigated the photographic method is much quicker than the electrical method .
* Owing to Mr. Moseley 's absen from England , the proofs of the stenographic report of his remarks have been corrected for him by Sir Ernest Rutherford .
An account of these experiments of Mr. Moseley has been given in two papers in the ' Philosophical Magazine , ' Dec. , 1913 , and April , 1914 .
The first slide shows the most convenient kind of -ray tube to use for this work .
It is simply a very large -ray tube of 3 litres capacity .
The different specimens which are going to be bombarded by the stream of cathode , in to excite their characteristic X-radiations , ars mounted on a movable anticathode .
These specimens may bs in the form of elements , .
alloys , or compounds .
In many cases I have used thin layers of the metallic oxides spread over a plate of nickel , for example .
The substances are bombarded by ths cathode rays , and the stream of -radiation mingled with the characteristic monochromatic rays goes through a fine limiting slit .
They pass out in a divergent stream and fall on the reflecting crystal .
The , as Brag has shown , of the from crystals of these monochromatic rays is that they are only reflected when there is a certain definite connection between the wave-length and the angle at which they strike the crystal .
Possibly only one of the different kinds of radiation will hit the crystal at such an angle that it can be reflected .
The reflected rays have this remarkable property that , although the incident beam is divergent , a sharply defined line of about the same breadth as the origiual slit is found on the photographic plate .
I have examined by this method a large nnmber of elements or their compounds .
This slide shows one of the earliest photographs obtained for chromium .
There is one extremely strong line and another much fainter line of about 10 per cent. greater frequency .
The spectrum shown for chromium is to be ascribed to\ldquo ; \ldquo ; characteristic radiation of that element .
For elements of atomic weight greater than silver , the spectrum of the " " \ldquo ; series of Barkla has been determined .
This spectrum consists mainly of three lines , , which are clearly shown in the spectrum of lanthanum , for example .
In this slide are given the collected results for all the elements that so far I have been able to investigate .
I thought that it was better at first to make a general rough survey of the spectra of as many elements as possible , rather than to go into very elaborate il .
We shall first consider the spectra of the series betweexl aluminium and silver .
On this diagram the abscissae represent the square root of the frequency , and the successive elements are arranged on the verticaI scale at equal intervals .
Suppose that we begin with aluminium .
This represents the very strong line of aluminium corresponding to the strong line of ohromium ) previously , and this the weaker lins of just a few per cent. greater irequency .
On coming to silicon , we get exactly the same phenomenon repeated .
Here is chlorine , of which at present I have only measured the strongest line .
Here are potassium , calcium , titanium , manganese , iron , cobalt , nicksl , copper , zinc .
Aftel ' a gap we come on to yttrium , zirconium , and so on .
You see that the square root of the frequencies of these strong lines plotted against the successive elements placed at equal intervals on this vertical scale fall on a straight ] , or rather a line that is straight at first and then bends over slightly .
At the latter part we are dealing with very high frequencies , so that whatever is giving this must be treated mathematically as an electromagnetic problem and not as the purely electrostatic problem which ] is usually taken to suit the lower frequencies .
The other fainter lines also fall on a very good straight line , as far as they have been traced , though for the high frequencies only the principal line has been investigated .
In making this diagram the elements , from aluminium upwards , have been arranged not exactly in the order of increasing atomic weight , but in the order of Mendeleeff 's pet.iodic table ; that is the order of the atomic weight except in the few cases where this clashes with the of the gradually changing chemical properties .
Here , for example , we have the order , chlorine , argon , potassium , calcium , instead of the atomic weight order , chlorine , potassium , argon , calcium .
If the atomic weight order were used the spectrum lines wouId be thrown completely off this straight line so we see that we are justified in taking them in the above order .
Here , again , the chemical order , iron , cobaIt , nickel , is chosen instead of atomic weight order , iron , nickel , cobalt .
It will be observed that the elements have all been spaced at equal distances .
This means in order to get some definite simple relation between ths frequency and the element we must ascribe to each element an integer .
This integer can hardly be anything else than the atomic number , the charge on the central nucleus , about which Sir Ernest Rutherford has been speaking , and the fact that the atomic number is about half the atomic weight fits in well with this .
We ascribe to aluminium the atomic number 13 , though in this point we are not supported by definite experimental evidence .
The atomic number of hydrogen is known to be 1 , that of helium is 2 , so that unless , which is extraordinarily unlikely , the chemists have missed an element before aluminium , the latter , the 13th element , has the atomic number 13 .
The resuIts show that no elements are missing between aluminium and silver , except for one gap left for a homologue of manganese .
Let us next consider the elements of the series .
Fortunately it is not very difficult experimentally to overlap from the series to the ; that is to say , for some elements both the series and the series be measured .
I might mention in passing that the characteristic of very long wavelength in this part of the tgram are rather troublesome to they havr such extremely small power of penetration .
A radiation of this kind is practically stopped by a centimetre of air .
Therefore , is necessary in investigating them to use a special vacuum spectrometer .
In these series spectra the strongest principal line is called , and , as before , it is the line of smallest frequency .
The points for successive elements lie on a definite regular curve which bends a little and is not quite a straight line .
The line , the next in order of intensity , also lies on a smooth curve which is very much bent .
The third line gives practically a straight line .
There is another line which at presenb I have only found in the region of the rare earths , although very probably it also appears elsewhere .
The results do not pretend to be in the least complete .
The line for platinum was originally called by Prof. Brag .
These two lines and were called before he knew that the line was double .
There are also some fainter lines of higher frequency not given on this diagram .
What I want to insist upon is the fact that each of the elements has got a spectrum extremely like that of its neighbour .
It is obvious that whatever is giving this spectrum is the corresponding system in the different elements and that its properties gradually change as a function of the atomic The continuity of these curves shows that the order in which we have arranged the elements must be correct .
I must say a few words about the way in which this order has arrived at .
In any doubtful instance it has been so arranged that the elements all lie on a straight line .
Chemistry always suspected that there was an owl ] homologue of manganese between moIybdenum and ruthenium .
Already in the ser.ies spectra it has proved necessary to leave a space for such an element .
If this element No. 43 exists at all , it ought to be very easily detected , because the frequencies of the two principal lines of its -ray spectrum can be predicted with certainty .
It is necessary also to leave a space for another homologue of anese between tungsten and osmium .
The only other doubtful the number of the rare-earth elements .
There has been a very large amount !
of extremely skilled chemical work done in separating these elements , but it is still a matter of some doubt whether some of them are really simple , or whether they are mixtures .
It has been shown that what we used to call ytterbium is a mixture of two elements now as -ytterbium and lutecium .
Vacant spaces are left for these two elements .
I find that between erbium and tantalum there are only four places to spare .
The spectrum of erbium was determined from two specimens .
The first was a specimen which Sir William Crookes had the great kindness to give me .
It gave faintly the strongest line of both elements , thulium I and thulium II .
There is no doubt that the two different kinds of thulium are exeedingly difficuIt to separate completely from erbium .
The erbia also contains traces of dysprosium and of holmium .
Another sample of erbia contained quite 50 per cent. of holmium , or rather of the element whose abomic number is 66 , which I take to be holmium .
Welsbach has actually separated thulium into two different substances , and his assertion that there is a third is based on rather more indirect evidence .
We will now look a little more closely into the question of the relation between the square root of the frequency and the atomic number .
From the curves we saw that the square root of the frequency was a linear function of the atomic number N. So we have the frequency .
If we take the strongest line of the series , we find that this constant A is numerically equal to where Eydberg 's fundamental frequency for ordinary line spectra .
I have written it in this form because it is then identical with that which one would for the principal line from the theory of Bohr .
The constant is equal to 1 .
Considering the principal line in the series , the constant is equal to about , i.e. , Thus for line while for the line Of course , one cannot determine the value of A closer than a few sr cent because we saw that the principal line of the series , when the square root of the frequency plotted against the number , was not quite a strai ht line .
It is very likely mereIy a coinciderlce that the value of A happens to come in that .
But I am intending to look for the series and the series , and , if they ex , to examine whether they also have constants of this kind .
I should like to point out that for the series the value of is considerabl lar er th the serie.s .
The complete difference between the grouping of the lines arger an the series and the series , and the fact that the constant , which I suppose go nes or to represent the action of the electrons in neutralising the force exerted by the central charge , is very different in the cases , suggests that the series .
corresponds to the vibrations of the insids ring , and the series to the vibrations .
of the next ring outside it .
This slide is of since it shows the spectrum of the specimen of erbinm which Sir William Crookes gave me .
The , 13 , and lines are clearly seen ; .
also a line of dysprosium , while the principal line of holmium is just visible .
It shows how these spectra can be utilised to determine the purity of the specimens .
The next slide gives the atomic numbers of all the elements , assuming that we are right in saying that aluminium is No. 13 .
You will notice that the numbers up to 79 have been put in from the evidence of ths -ray spectra .
After gold , No. 79 , we assume ihat the same Iaw continues to hold for the two elements , mercury and thallium , which I have not yet investigated .
After thallium we get into the region which Mr. Soddy has been studying , and , from the evidence given him , we can deduce the atomic numbers of all the other radioactive elements up to anium .
You see that there is one more element missing , a homologue of chlorine , and also a homologue of sodium , but it says a good deal for the extraordinary skill and indust of the chemists , that in the whole list there only seem to be five unknown possible elements .
Mr. F. SODDY : I have listened with very great interest to ths discussion opened by Sir Ernest Rutherford upon the structure of the atom , and though of conrse I am not competent to discuss the nuclear atom , if I may give it a name , critically from the mathematical or the physical side , I may say that generally in the broadest way I am in full agreement with that model .
I have certainly found it very helpful in my particular field of work in radioactive investigations .
Prof. .
more than ten years ago , showed that an -particle could go through all the atoms it met in its path without suffering any appreciable deflection .
It is only a particle moving at these uItra-material speeds , speeds which are about one-tenth of the speed of light , such as the -particle , which can penetrate the atom , and , as the genius of Sir Ernest Rutherford has shown , it may tell us something about what it is that it has passed through .
The fact that the law of Brag is not actually obeyed and that there are these rare single scatterings , to use Sir Ernest Rutherford 's term , I think , without any theory whatever , does .
show that we have to deal with a nebular atom with a nucleus .
I am not very friendly to the general idea of the electrical theory of matter ; but , without any particular assumption as to the nature of the atom , I think that the resuIts of single scattering as shown by Prof. Rutherford do undoubtedly prove that we have to deal with a nebular atom which has a hard point in it , a hard scattering point which can deflect a heavy atom travelling at very considerable velocity .
I content myself by merely saying that , of course , the electrical theory of matter has now undergone very profound change , because in the Thomson atom there was an attempt made to preserve stability , an attempt which was not altogether successful .
But with this new atom we have got beyond any such attempt to preserve stability .
Because it is perfectly obvious that this nuclear atom , tever the nature of it may be , cannot obey the ordinary laws of electricity .
You could not by any known means concentrate into a space , 10-13 cm .
in radius , 100 atomic charges of the same sign .
There is no known way of doing it .
You have endowed electricity with the attributes of matter , rather than explained matter in of electricity .
Passing to my own line of work on this subject , let me iefly describe what led me to the view that there exist in the series certain elements which had identical chemical properties .
It was before that there were certain radioactive elements , radio-lead , for example , which extraordinarily closely resembled other elements , such as ordinary lead .
My first practical experience came in connection with meso-thorium .
The chemical nature of meso-thorium having been kept a technical secret by its discoverer , I had to work it out for myself , and I was amazed to find that , though of a different atomic weight , it yet possessed properties chemically identical with those of radium , and by no known process could you separate those two elements once they were mixed .
The work has been carried on by 1Ir .
Fleck at Glasgow , who has investigated all the radioactive elements in turn , to see , firstly , whether they resembled any known element , and , secondly , whether they were separable from that element .
He showed that every single member of the radioactive series , which has a period of life above a minute , is identical chemically with one or other of the last 12 elements of the periodic table from uranium to thallium .
This is a very important point , because you have a new means here of determining a type of complexity of matter which a chemical analysis must necessarily overlook .
Suppose that you have a radioactivs sequence such as , and F. If we take a common case , , the primary element , may be chemically identical with with , and with , and by no chemical process would you be able to determine that six elements were .
Chemical analysis will separate three pairs of substances , namely , A and and and F. But if it happens that the periods of average life of the various members of the series are not long in comparison with the allotted span of three-score years and ten , , and in absence of their respective parents , , and , will disappear with lapse of time .
But A will , in time , change into into , and so on .
Hence all six members can be severally distinguished .
Radioactivity provides , therefore , means of distinguishing a complexity among elements in the state of change which is not within the range of chemical analysis .
It was in this lvay that Sir William Ramsay and Hahn discovered radio-thorium .
IRadio-thorium , as we ow now , is non-separable from horium .
In fact , these tors had separated out nleso-thorium from the mineral unwittirlgly .
As meso-thorium was not then , they naturally assumed that they had separated the radio-thorium from the mineral .
We do not know any process at present capable of removing radiothorium from thorium .
They really had separated meso-thorium from the mineral , which , in course of time , ohartged into radio-thorium .
I would like to have my lantern slide .
I showed these to the Royal Society over a year ago .
I need not spend time upon it .
It shows the radioelements arranged in the periodic table according to the chemical results of Fleck and myself , and a great number of other irlvestigators preceding us .
Every vertical row consists of elements occupying the same place in the table , which are not separable from one another by any known process .
Such elements I call isotopes .
Two years previously I had shown that the expulsion of an -particle shifts an element from one place in the periodic table to the next place but one .
That *See , Chemistry of the Radio-Elements , ' Part II , p. 3 .
was followed up by my old student , .
A. S. Russell , and at the same time by Herr Fajans in Germany , who supplemented that rule and showed that the expulsion of a -particle shifts the element one place in the opposite direction .
Hence the expulsion of one and two -particles must necessarily bring the element back again to where it started , and give yott an element completely dentical with that which you began with .
Owing to the existence of intermediate members of entirely chemical you can get the radiothorium and thorium separately , when by chemical analysis such a thing would be absolutely impossible .
The result of the experiments of Messrs. Russell and Ross , and of Profs .
Exner and Haschek , of Vienna , is to show that the spectra of these isotopes , as I call them , are very probably identical .
They showed that it was not to get any new spectra lines in a very strong iomum prepa .
The spectrum was purely that of thorium .
Although we do not kuow that the chemical elements are not homogeneous in other parts of the periodic table , the presumption is that they are probably not homogeneous .
In the last part of the periodic law one has a definite proof that there are 40 elements crowded into the 10 places which are occupied .
The atomic may over a considerable number of units from 206 , for instance , in the place occupied by the element lead , to 214 , 8 units , so there is a possibility that the same occurs in other parts of the periodic table too .
Since the expulsion of an -particle with two positive corresponds with a shift of two places in one direction , and the expulsion of a -particle a single negative charge.corresponds with a shift of ons place in the opposite direction , it follows without any doubt at all that the successive places in the periodic table correspond with unit differences in the net value of the nuclear charge , so far as the last 12 places at least are concerned .
Thab idea has been followed out further by Mr. Moseley , who , I understand , has new results to bring before us to-night .
There are a great many problems which have not yet been solved .
It is probable that the ordinary lead of the chemist is not homogeneous .
If it is derived from radioactive changes it is a mixture .
There has not yet been time for thali point to be decided , but we have been at work on it for some time .
I regard the discovery of meta-neon and neon in the atmosphere by Sir J. J. Thomson as a brilliant confirmatio1l of this point of view .
These elements , according to the results of Mr. Ashton , appear to differ in atomic weight by 2 units , and yet appear to have the same boiling point and spectrum .
They ars not capable of separation by fractional condensation in cold charcoal , but are separable by fractional diffusion methods only .
The model on the table of the periodic law I think sums up the knowledge we have derived from radioactive changes succinctly .
It is based on the figure-ofeight spiral arrangement of Sir William Crookes , now over 25 years old , and on the fact that the rate of change of properties in passing from place to place attains a maximum at the zero and a minimum at the eighth group .
We cannot take the simple symmetrical idea that then sufficed .
It illustrates three points which , I think , have to be explained on any theory of the structure of the atom .
Firstly , we begin with two simple single of the spiral , or short periods .
Then at the element , titanium , the course of the evoIution changes * See ' Chemistry of the Radio-Elements , ' Part II , p. 11 .
gives us the double spiral figure .
This proceeds till complete double turns have been run through .
Then we have a totally different type of evolution coming in , with the commencement of the rare-earth group .
All the elements are identical in valency , though they differ , of course , from the isotopes whehI spoke of ; in that between the successive elements there are distinct , though exceedingly slight , differences of chemical charaoter , sufficient enable their separation by chemical methods to be effected .
The sequence then follows exactly as though the rare-earth elements had not been interpolated .
The final part between bismuth and uranium is almost all new , but it rsproduces in exact detail the earlier sequence of events .
From lead to thorium is the exact analogue of the early sequence from carbon to silicon .
Prof. J. W. NICHOLSON : I think that Prof. Rutherford has made it clear that the nuolear atom is , as a matter of fact , the only basis upon which profitable discussion of the constitution of the atom can really be carried on .
The main theory , as Prof. Hicks has just indicated , is the magneton theory .
It seems that just recently in the course of a discussion it has been shown that the magneton theory and the nuclear theory do probably amount to exactly the same thing .
This has been indicated very clearly by Dr. AlIen , who , I observe , is here to-day , and who , perhaps , may speak about it afterwards .
He has shown that the magneton can be represented as a difference effect which takes place with rotating electrons , if we suppo.se that the nucIeus is rotating as well , with , however , specified angular momentum .
The principle of the constancy of an momentum gives a possible explanation of Planck 's constant , that is to say , if this constant is an atomic constant at all , and not a constant of atomic aggregates .
It first came in in connection with the coronal spectrum , which can be represented mathematically by the free vibrations of a very simple model atom in its neutral and charged configurations .
( Illustrating .
) That is an atom which has a nucleus of charge 5 and five electrons .
This angular momentum principle was taken up by Dr. Bohr and is the real basis , I think , of the success of his theory so far as it has had success .
A hope has been expressed by Prof. Rutherford that I would make a few remarks about Dr. Bohr 's theory .
I have been examining this theory for some little whils ; in fact practically ever since it first appeared .
There are several points in which it will work , and others apparently in which it will not work without very serious modification .
In the first placs thers is the question of the theory so far as it reIates to the more chemical nature of the elements , that is to say , to the arrangement of the electrons in the atom in a series of concentric rings of varying radii , the X-ray , for example , being supposed to come from an innermost ring .
On the older electrodynamics it is possible to show that we cannot have , under any circumstances , a set of rings in the same plane rotating round a single nucleus .
It is not possible to satisfy the conditions .
We can have , of course , electrons rotating in other ways than all in the same circular ring .
We can have them all describing ellipses , for example , with the major axes of ellipses at equal angles all way round the nucleus .
But it is not possible to have any approximation to concentric circular orbits on the old electrodynamics ; nor is it possible , as can be proved rigorously\mdash ; I mean by strict mathematics\mdash ; to have them on Dr. Bohr 's theory , either , in its present form .
On this theory at present the so-called steady configurations of the atom are to be determined by the ordinary mechanics .
But it is only when we come to the question of the changes of the atom from one set of stationary states to another , or the vibration of the atom , that the ordinary dynamics must be given up .
In obtaining the formula of the hydrogen spectrum , Dr. Bohr is compelled to suppose that the attraction between an electron and the in ths case of the simple neutral hydrogen atom is according to the ordinary inverse square law which it would if the electrons were not actually bound in an atom .
If we examine what the spectrum of any other atom would be on these premises , that is to say , take any case which Dr. Bohr has not taken , we cannot obtain the experimental resuIts .
But taking first the hydrogen atom in the ordinary neutral state , that particular atom also leads to another spectrum in the extreme ultra-violet , which has recently been found and which rees very well indeed with the theory .
In other words , the hydrogen atom certainly does appear to give all the lines that we can get from a certain double series .
This includes Balmer 's formula , Ritz 's formula , and a Schumann series .
The formula is , where is Rydberg 's constant , and and can take all integral values .
In the case of helium , when thers is only one electron approaching the atom , we get the same series with half instead of whole numbers , and one immediate dednction , among others , is that the entire hydrogen spectrum is also given by the helium atom when it has a posibive charge .
Actually there is a slight correction to be made in , because when you have only one electron , it drags the positive charge out of its ordinary position .
The two series for hydrogen and helium are not really quite the same , but they are very nearly .
If we take the same premises and try to get the rest of the helium spectrum\mdash ; that is to say , the dinary helium spectrum\mdash ; the attempt fails altogether .
We get other series of this kind if we work out the varying stages according to the theory given by Dr. Bohr .
If we try to get the spectrum of a charged hydrogen atom or a hydrogen molecule , the same thing takes place .
We can get nothing approaching the facts at all , and we must bear in mind that we have an entire secondary spectrum of hydrogen , or what is now generally believed to be due to hydrogen , with about 400 lines , which has yet to receive explanation by any theory of the structure of the hydrogen atom , in addition to those theories which have been mentioned .
I have examined in what direction we can modify this theory in order to get such spectra .
I find that there are two ways .
We can either make a different specification of the angular momentum\mdash ; but that is not sufficient in itself\mdash ; or we must suppose that , although the electron and the nucleus exert the usual force on each other , electrons bound in an atom must not exert the same force on each other as they normally do .
I think that , in dealing lvith a subject like this , we are , perhaps , entitled to make any tentative assumption we like , which Ieads to an experimental formula , in order to endeavour to throw some light on the subject .
I find , however , that no inverse law of any kind will lead to anything approximating to the ordinary spectrum of helium at all .
The only way in which any progress seems likely in connection with ordinary spectra is to suppose either that there is no force whatever between electrons bound in an atom , or that it is according to some power which is not an inverse power .
When we come to the -ray spectra some very interesting resuIts appear .
The first point that I wish to emphasize is in reIation to the ordinary deduction first given by Mr. Moseley , that 's theory is in agreement with his resuIts on the -ray spectra .
When we examine Bohr 's spectral formula , we get : .
is a constant depending on the number of electrons , that is to say , in addition to that factor in , the factor comes in .
This has been overlooked in the treatment of the question so far .
There is only one way of removing , and that is to give up one of Dr. Bohr 's fundamental postulates , which is to the effect that when radiation is given from an atom only one quantum must go at a time .
If there are electrons , we must suppose that every one of those electrons sends out a quantum .
If we do this , we can get the on the other side of the equation , and it may be cancelled .
That is the first proviso we must make if we want the X-ray spectra to fit in with the theory .
If Mr. Moseley 's rings are impossible , then his supposed ring of four electrons round the nucleus will have to go too , except in one circumstance , namely , that there is no force between the bound electrons in the atom .
If there is no force in the bound electrons in the atom , then we can proceed a little farther .
The constant disappears .
We get the frequencies of the -ray spectrum proportional to the square of N. ( Illnstrating .
) Mr. Moseley 's results do indicate so closely this propor- tionality to a square that I think it would , perhaps , be er to accept it as a fact rather than ) pose that there is a constant there which happens to be equal to 1 in the case of calcium and some of the other elements .
Here , vever , we have the difficulty that if represents the atomic number the for calcium comes down to 19 ; that is to say that if the theory of the atomic number being equal to ths place of the elements in the table is correct the chemists have not found exactly the right number of elements between hydrogen and calcium , but have found one too many .
However , I do not want to dwell on that point now .
If this absenc-e of force between bound electrons is a fact , isotopes will have the same spectrum .
There is one other point on which I would like to make one or two remarks .
I referred to the question of the coronal spectrum , which can be explained by the vibrations of an atom consisting of a nucleus of 5 , and 5 electrons .
This atom I do not take to be an ordinary terrestrial atom of similar constitution .
I take it to be an atom with a simple nucleus .
I think that the nature of the binding of the eIectrons to the nucleus must indicats some change in it from the form of a terrestrial atom .
If we examine the spectra of nebulae in the same way , we get the same result with respect to a system of four electrons Qide a nucleus of four .
We can explain a large part of the nebular spectrum in this Recently I think I have managed to show that we can explain practically the radiations of nebulae by the other systems of this simple kind ; that is to say , we have the very beginnings of matter there ; and the subsequent development of terrestrial matter will have to take place by some modification of the intsrior of the atoms .
The evidence for this modification is , that when we come to the stars which astronomers regard usually as the earliest type of stars , namely , the stars close to the nebulae in stellar classifications , we find that they contain lines which can be identified except hydrogeu and helium lines .
They contain , however , about forty lines which can all be grouped together under one comprehensive formula which is exactly like Balmer 's formula , except that the constant 4 is ] aced by something else , and there is a series of limits with a frequency difference outside .
The constants in these formulae are certain constants we can calculate for the original system in the nebulae with four electrons and the nucleus of four which seems to be the recognised system there .
There is evidently some relation between the origin of those lines ind the origin of the lines in the nebulae .
Moreover , if we trace through exactly the same relation for the subsequent products of another atom\mdash ; that is to say , the atom with a simple nucleus ( whatever that may be ) and one electron\mdash ; we get exactly Balmer 's formula .
We get the limit ight .
We get it right to five figures on the supposition that this explanation of the nebular spectrum is Apparently therefore hydrogen does contain only one electron .
Moreover , if we make the same calculation for an atom with two electrons we get this as the chief line .
( Illustrating .
) I do not deny that there are probably other series .
We get the wave-length 10832 as the principal line , and strongest line in the helium f$pectrum is 10830 .
I think that from this empirical method of there is very strong evidence that the supposition that hydrogen has a nucleus of 1 and helium of 2 is correct .
If we go any farther the same thing does not happen .
We cannot show in the same way that the nucleus of lithium has a charge of 3 .
There is one little difficulty which I.may , perhaps , refer to as regards the spectra .
Prof. Strutt recently has soms experiments which seem to a view expressed by Lenard some little while ago that the systems whose give the principal sel'ies and the subordinate series are different .
Lenard 's view was that the principal series came from neutral atoms and that the subordinate series came from the various positively charged atoms .
If helium has only two electrons we cannot expect a spectrum when no electrotls are left at all .
In that case the subordinate series of helium must be the spectrum of the atom which , according to Dr. Bohr 's theory , us what are called by astronomers the Pickering lines .
The line 10830 , of the principal series , by our method appears to be in agreement with Lenard 's view , for it comes from a neutral atom .
There will be these points uItimately to be reconciled between Dr. Bohr 's view and this more empirical view which I have ventured to bring before you .
Prof. W. M. HICKS : As there is not time to say very much , I will go at once to the one or two points that I wish specially to draw attention to .
The first is in reference to Mr. Soddy 's theory of the existience of elements different atomic weights , and yet having the same chemical properties .
I think that shall all agree that the evidence in favour of that is so strong we may take it as practically , or at any rate provisionally , settled ; but a difficulty ises in connection with the spectra .
He assumes that the spectra of these different substances with different atomic weights , but belonging to the same place , will be identical .
Certainly in favour of that view the result of the attempt to determine the spectrum of ionium seems to lend some weight , but there is a difficulty which it seems to me we cannot get over , namely , that the spectra of the elements depend very definitely on the atomic weight and not on the number .
It is not simply a rough connection .
It is a connection so close that if you take the atomic weight to be 1 or 2 in 3000 different from the weight as determined by chemists , then the relation falls out aItogether .
Here we have cases where the atomic weight differs by 3 or 4 units .
It is difficult to see how you can expect to get the same spectra .
An explanation may lie in the possibility that we never have the single element in these cases , so that , for instance , in the spectrum of we do not get the spectrum of pure thorium , but that of a mixture of thorium and ionium and all their isotopes together .
All these elements have an enormous number of lines , and their spectra have none of them been investigated , so we do not know whether we have a of the spectra of several isotopes when we take , say , the spectrum of thorium , or not .
Of course there is another supposition that we might take , namely , that the isotope with the largest number of electrons is the one which gives the ordinary spectrum , and that these other elements which are brought back by the emission of -rays have no spectra at all .
But it seems to me that the most likely thing and the thing that we have to Iook for is ] the observed spectra are really composite .
The other point which I wanted to raise was ono of a different kind .
If we approach the constitution of the atom from the spectral sids we are led to think more of the magnetic qualities of the atom than of the electrical .
You must have magnetic fields , in anything which produces ordinary spectra , at least those which show spectral series .
This has been pointed out by Lord Rayleigh and has been applied great success by Ritz .
Not only so , but we know from Weiss 's experimental work that these little atomic magnets do really exist .
Whether they are something different from electrons , or are produced by rotating electrons , is a matter for further consideration ; but the fields are so strong that they must be taken into account .
The measured amount of their magnetic moments is of the order .
Although the mathematical difficuIties may be great , I think that if the theory of the deflection of an -particle by one of these magnetons were worked out it would be found that deflections quite comparable to those observed would be found , and that too by particles which do not to go so close to the central nucleus as in Prof. Rutherford 's theory ; but of course whether the law of scattering is the same as what he has observed or not , we do not know .
That is a matter in which the mathematical theory would have to come first .
Some , rough calculations I have made show that or -particles passing in the equatorial plane of a atomic magnet do , as a fact , get scattered in all directions by such an amount as to lead to an expectation that the law wouId be similar to that which Prof. Rutherford has observed .
Prof. SILVANUS P. THOMPSON : May I be allowed to put a point 1 Prof. Hicks has referred to a feature which apparently has escaped the attention of almost all the older workers on this subject , namely , that there must be ma netic foroe these hypothetical atoms , if they consist of electrons in motion .
He has referred gne oroe also to the brilliant work done by Prof. Weiss in his err lnvestigation of the possible atomic magnet or .
It is worth while to put that from another point of view , to show one of the additional omplications that this consideration introduces .
What Prof. Weiss has done is this :Taking various magnetic substances , and magnetising them in the strongest possible field , so as to produce as near an approach as possible to magnetic saturation , he has found that there is apparently a definite magnetic moment for the atom of each of these .
Putting it into the usual quantity of gramme.molecules , .
of iron appear to have , in the limit when they are magnetised completely , a magnetic moment of 12360 ( C.G.S. units ) ; 59 .
of nickel , the corresponding quantity , appear to have 3370 units of magnetic moment .
Those two numbers have , as Weiss points out , exactly the relation numerically to one another of 11 to 3 .
If you take 59 .
of cobalt in the same way and express its magnetic moment in the same terms , the unitary number that corresponds to it is 9 .
If you take 16 .
of oxygen , the corresponding number is 7 .
That is to say , the atoms respectively of iron , nickel , cobalt , and oxygen contain these numbers of magnetons : iron , 11 ( or , in other circumstances , -iron apparently will contain 12 ) , nicke13 , cobalt 9 , and oxygen 7 .
Now , what I want to point out is this : we accept the argument of Weiss on this point , we must consider the atom not only as a nucleus with a certain electric positive charge in it , and electrons , negative units , grouped alound in rings or some other configuration , but its configuration must be such as to have these elementary quantities 11 , 3 , 9 , 7 for the four elements in question .
Arrangements have been proposed of models of atoms having so many units of positive electricity grouped with so many electrons of negative electricity .
Have we there any basis whatever for supposing that these unitary magnetic numbers can be accommodated therein ?
Our model atom of oxygen , however many positive and negative aggregate units it is supposed to have , must possess a configuration such that it shall have seven magnetic units in it .
This is a somewhat novel point of view , but apparently it is a necessity from the facts .
It adds , no doubt , to the difficulties of imagining a model that will work , but a model that will not work magnetically , but only mechanically , is only half a model .
Dr. H. S. ALLEN : Like Prof. Hicks I have been brying recently to work oul the question of the scattering of the -particles by a magnetic atom .
I find , ho wever , that there is one ) oint of difference between the electric and the magnetic problem , which I think is crucial .
In the case of the nucleus Sir Ernest Rutherford has that the scattering varies inversely as the fourth power of the velocity .
Taking the magnetic atom and dealing with the case in the equatorial plane , I think that the scattering varies inversely as the square of the velocity .
So there is this important point by which we may distinguish between the two views .
I am not quite sure whether the experimental evidence obtained so far is really decisive .
I think that if we were to take a magnetic atom which also possesses electrical properties we might find a relation somewhere between the square and the fourth power which possibly might fit the experimental facts more closely .
With regard to magnetons , I want to mention one point very briefiy .
We may express the angular momentum of an electron as , supposing that we have an electron travelling in a circular orbit of radius , with angular velocity According to Bohr 's theory the angular momentum is equaI to .
The magnetic moment of an elementary magnet , which would be equivalent to the revolving ( 8001 ) electron , is equal to .
Comparing these two , we see that it is possible to express the magnetic moment in terms of Planck 's constant .
If we out the result , using the latest value for the constants , find for the magnetic moment the value .
Now , taking the latest for the magnetic moment of the magneton we get the number .
The former mber , then , is five times the latter , that is to say , the single electron revolvin in a circular orbit is equivalent to five magnetons .
That was pointed first by Mr. Chalmers at the recent meeting of the British Association .
In order to arrive at the magneton itself , I suggested in a letter to NalUTe , ' a few weeks ago , that we might take the case of a revolving sphere of positive electricity .
It is easy to show that , if we have a sphere carrying a charge times , the magnetic moment of the revolving sphere is , where A is the radius of the sphere and its angular velocity .
If we assume that is equal to , the magnetic moment is equivalent to magnetons ; so that , by taking magnetons for the ring of electrons and tons for the revolving sphere , have a way of building up various numbers of magnetons according to the direction of revolution of the sphere and the ring .
Prof. Rntherford , I suppose , would prefer to make the sphere of negative eleutricity and pnt a positive nucleus in the centre .
In that way we might be able to reconcile the two points of view .
of SIR ERNEST RUTHERFORD.\mdash ; As the hour is so late I will only delay you for a few minutes .
I must first of all thank the speakers for the kind way in which they have treated me in this discussion , and I think there are not many points which call for a special reply .
I quite recognise , with Prof. Hicks , that there must be an atomic weight term connected with the frequencies of vibration of the atom .
On the theory I have outlined , it is to be anticipated that this term should be very small , and it is difficult to account for the magnitude of the effect observed experimentally by him .
In making the statement that elements of the same nucleus charge should have the same spectra , I am quite aware that this can be only approximately the case .
There should always be a small in the frequency of vibration due to the mass of the nucleus , and no doubt the magnitude of this term will depend upon the relative masses of the two atoms , which have the same nucleus charge .
According to our present views , however , this effect should be very small , and for a heavy atom should not amount to more than 1/ 100 of an Angstrom unit .
In my remarks I did not discuss at all the very thorny question of the outer distribution of electrons .
There is no doubt that in a heavy atom this distribution is very complex and will be difficult to calculate with certainty .
I am very glad that Prof. Nicholson has given us his views on the origin of the spectra , and was very interested in his discussion of how far the theories of Bohr are applicable , and how the results given by Mr. Moseley can be interpreted .
Even supposing that our fundamental assumptions are right about the nucleus and the importance of the nucleus charge , I think it will be some time before we shall be able to enter with any detail into the very complicated question of the electronic distribution of a heavy atom .
There are so many possible modes of arrangement of the electrons , and there is always the uncertainty that we do not know completely the laws which govern the influence of one electron on another .
Obviously very many possibilities will have to be considered before we can hope to obtain a satisfactory explanation .
Any structure which is assumed must explain not only the general chemical and physical properties and the spectra of the elements , but also the magnetic ertieS of the atom , to which I am glad attention has been directed by Prof. Silvanus Thompson and Dr. Allen .
I am afraid that I have already taken up too much of your time in this discussion .
I personally take a very interest in this snbject , and to-day I have tried to give you briefly my general views and the theories on which many of us are working to-day .
I am glad that we have had such an in terchange of views on soms of the main problems of atomic structure , and that there is com- paratively so much agresment in regard to the central point of the atom .
E. R. * [ Added June 19 , 1914.\mdash ; It must be borne in mind that the magnetism of iron and nickel are very much affected by their physical and chemical state and their combination with other substances .
lt thus appears probable that the intense magnetie effects of iron and nickel are in some way connected with the arrangement and motion of the electrons near the surface of the atom , and are only indirectly connected with the interior strucbure of the atom.\mdash ; E. R. ]

