]\gt ; [ 459 ] XI .
Enquiry into the Variation of Angles observed in rystats ; especially of Potassihm and B.y HENRY A. MIERS , D.Sc .
, M.A. , F.R.S. , UP .G.S .
, .
P. C.S. , Professor of the of Oxford .
Received ]iead M ) CONTENTS .
PAttT OF ALUM ) VlClNAL 1 1 .
The variations of angle previously observed . . . . .
460 2 .
Object of the present investigation . . . . . . . . .
} 3 .
The goniometer . . . . .
464 4 .
Purposes for hich the goniometer can be use . . . . .
466 Etching of calcite with HC1 . . . . .
467 ( B. ) rallel growth of with calcito . . . . . . . .
468 5 .
The hedron angle of . . . . . . . . .
469 6 .
The natnre of the vicinal faces on potash-alum and anunonia-alum . . . . .
7 .
The vicinal faces on sodium chlorate , zinc sulphate , and magnesium sulphate 483 8 .
The effect on alum of ting the solution . . . . . .
487 9 .
Some possiblo causes of the Yariations . . . .
\mdash ; THE ATION o THE 1N TACT wlTn Gr.0W1NG CRYSTAL .
A. Preliminary experiments . . . . .
493 B. The refractive indices of supersaturated solutio1ls . . . . .
; ( 1 .
) Potash.alum . . . . .
497 ( 2 .
) Tartaric acid . . . . .
601 ( 3 .
) Sodium . . . . .
504 ( 4 .
) Sodium nitrate , . . . .
C. The refracve index conccntratio of the tion i with a growing . . . . . . .
( 1 .
) .
, .
Specific vity of in contact of -alnm ) Sodium . . . . . . . .
) ( 3 .
) Sodium . . . .
, . . . . .
PART III.\mdash ; THE ACTIO or Interpretation of the fot.egoing resuIts .
A possible c$use the tion otngle in and other Ct.ystals . . . .
: . . . . . . . . . .
xperiments npon -alum crystallising in tiott . . . . .
PROFESSOR HENRy A. MIERS : AN ENQUIRY INTO PART I. ( 1 .
) ANGLE RVhD .
ALL who been practically aged in goniometric measurelnents are well that in crystals of one and the same substance small variations of angle are usually encountered even when the crystals are taken at the same time from the same solution .
The differences are sometimes so small that they are only appreciable by a goniometer which will read accurately to 1 minute or less , and very often the crystal faces themselves are too irregular or too imperfect to yield sharp images of the slit such that they can be adjusted within , say , half a minute .
But eyen if account be taken only of those faces which are truly plane and yield sharp reflection of the collimator slit , the individual variations may amount to 5 minutes , and the nueasurements of the same angle made upon , say , 20 good crystals of the same substance , may differ to this extent .
I have studied* an example from the mineral kingdom in the case of the very beautifully crystallised Proustite , where the angle between perfectly smooth and plane rhombohedron faces yaried between and upon tl , same crystal .
It is customary to eliminate these variations , as well as errors of reading due to the imperfection of the faces , by measuring a considerable number of crystals and taking the mean of corresponding measurements .
This process appears to lead to very satisfactory results , for angles may sometimes be obtained which are consistent even within half a minute of arc ; that is to say , the lengths and mutual inclinations of the crystal axes having been calculated from some of these angles , the remaining angles are calculated on the assumption that they belong to planes whose intercepts are certain rational sub-multiples of the axes , and they are then found to agree very closely with the means of the observations .
Two { tions are here made : ( 1 ) that the individual variations are just as likely to give angles which are too large as angles which are too small , and ( 2 ) that the faces do obey the law of rational indices .
Even when the calculations are made by the method of least squares , these assumptions still underlie the process .
The small individual variations are , in fact , regarded as irregular and of secondary importance , and are practically ignored ; if they be thus left out of account , it cannot be denied that within narrow limits the angles of most crystals do approximate very closely to those required by the law of simple rational indices .
Occasionally , however , the deviation is too great to be treated in this way , and the direction of the face can * Mineralogical Magazine , ' vol. 8 , 1888 , p. 49 .
THE VARIATION OF ANGLLS IN no longer be expressed by simple rational indices , but only by very high numbers .
Cubes of fluor , for example , will often yield no reflections corresponding to the true faces of a cube , and , therefore , do not possess angles of ; but each apparent cube face really consists of a very flat pyramid whose adjacent planes are inclined to ] other at the angle .
These faces would intersect the of the cube in the ratio 1 : , and belong to the form .
uch " " vicinal faces\ldquo ; are encountered upon of most substances ; } , ] they by the solvent action of the lnother-liquor ; but they may certainly also appear during the growth of the crystal when it is not bein redissolved .
The vicinal faces themselves are sometimes as perfectly and smooth as any on the crystals .
In the case of fluor , the presence of the ) of foul such planes in place of the cube face shows that they not merelv due to the distortion of cube face , but to a different form .
But even when a face with high dices occurs upon a crystal as a single isolated plane , it is generally regarded as vicinal face :and either as due to the etching action of olvent , or as replacing with indices owing to the ) eration of some unknown cause .
Whether the vicinal obey the I of indices at is not certain .
Planes to which it is to attribute high indices are usually called ' ' vicinal and are regarded as something different from ordinary faces ; many careful measurements of individual vicinal planes have been made in order to determine their indices , whereas it is the custom to eliminate variations in the angle between what appear to be faces of simple forms .
Sometimes , however , these variations are so large , eve in the case of what appear to be quite simple forms , that they can scarcely be ignored in this mallner .
In most of the crystalline systems the theoretical allgles not known , but have to be calculated from some of the observed angles , which may be themselves liable to these variations , so that it is difficult to say whether the variations really irregular .
But in the cubic systeln , by virtue of its symmetry , angles known absolutely , and it is possible to coml ) the measured angles with the theoretical values .
Crystals belonging the cubic system ppear , howevel , to be liable to the same variations of angle with those of other systems , and also exhibit vicinal faces .
Now the whole value of the law of rational indices , which is the foundation-stone of crystallography , rests upon the permanence of angle .
The faces of an octahedron belonging to the cubic system are inclined to each other at ( thd , similarly , the of the form ( 112 ) , of which any face cuts two of the axes of the their extremities and bisects the third , is If the octahedron and icositetrahedron faces do not make these angles , respectively , then the law of rational indices is only a first approximation , and there is some disturbing influence has yet to be investigated .
HENRY A. MIERS : AN ENQUIRy INTO It must be observed that two interpretations are possible : ( 1 ) That the faces really simple indices , but ore liable to irregular ; ( 2 ) tlJey have ot really } ) indices , but vicinal Thu , oct edron ] ( is found to ) not but , would on tho first octahedron of which one or more faces are distorted their position by some unexplained cause ; on the second interpretation they would not be faces of ) octahedron at all , but a vicinal form having the indices , of which each ) inclined minutes to the octahedron face with whioh it nearly coincides .
In the simple case here contelnplated it ought to be quite easy to distinguish between the two interpretations , for the form ( 65.65.64 ) consists of , not 8 , but 24 faces , and each face of the apparent octahedron should yield , when measured on the goniometer , not one image of tlJe collimator slit , but three images , due to three facets , each by 25 from the octahedron face .
carelul upon the angles of cubic have been made by several investigators .
in 1878 an tion on the variations of crystal angles , in which he found that angles which were known from the symmetry of the might differ by 30 minutes from the theoretical value , or might agree with it to within 1 minute .
He came to the conclusion that the variations of angle are related to the existence of optical anomalies , and that in the cubic systelu those which are birefringent exhibit these variations , whereas those which are isotropic do not .
A natural deduction would be that the ) served variations are the result of strain .
Some years later problem was proposed for a prize essay by the Philosophical Faculty of the University of Marburg , and the inquiry was limited to crystals belonging to the cubic system ; candidates were directed to measure isotropic and birefringent crystals of the same substance with the view of .
whether any such relationship exists .
The essay , .
BRAUNS , published in 1887 .
He carefully measured octahedra of lead nitrate , of , and of ammonia-alumina-alum , choosing both isotropic crystals , and came to the conclusion that there is no difference between the angles of isotropic and birefringent cubic crystals .
Of 120 measured angles ( on 15 crystals ) 86 gave a deviation of 5 minutes or less , and 63 deviation of more minutes .
The largest deviation from the theoretical value was 19 ' ; the largest deviation , for the angle as measured between faces which gave very perfect images was 13 ' 20 in the case of a crystal oflead nitrate .
BBAUNS further made an interesting suggestion regarding the cause of these 'Sitznngsber . .
Physik .
Med. Soc. zu Erlangen , ' Heft 10 , p. 59 .
Neues Jahrbnch , ' 1887 , ( 1 ) , p. 138 .
THI OF RVED IN deviations ; the crystals of lead showed a certain ularity in the angles ; they had all grown lying on an octahedron IlOW if the obtuse angles made by the upper face ( 111 ) with the three adjacent faces be compared with the acute angle which it makes with the three other faces , it is found that the average value of the is less than and that of the latter is greater than BRAUNS drew the conclusion that the action of gravity is the disturbing cause ; this is not perceptible in spinel and alum , but does come into evidence in the denser substance , lead nitrate .
The inyestigations of BRAUNS are the only precise and direct attempts to determine and explain the angular variations of cubic crystals .
They lead to the conclusion that the planes of the crystal are octahedron faces , distorted by some cause which in the case of lead nitrate appears to be the action of gravity .
The faces were all sufficiently good to yield measurelnents 1eliable to 1 minute , and the author makes no mention of multiple images , so that they were all presumablv single plane , surfaces , and vere , therefore , as ) vicinal bnt ctalledron THE One serious difficulty ttcn 1 ) to study the problem by the methods of crystal measurelnent .
Where several octahedron Cl.ystals , for example , are taken from a solution and found to give allgles which not the theoretical angles of the , it is impossib]e to ascertain ] face of the one crystal corresponds exactly to ally of another .
It occurred to me that the only way of studying changes in angle , of ting kindred problenls , is to ascertain how far the val.iations are constant for one and the same crystal its growth .
For this purpose it will not suffice to a crystal out of the solution and measure it , and then to re-immerse the crystal , allow it to measure it again , for one is whether any faces of the parallel to those of the shell by which it becomes enyeloped ; it ; necessary to a plan by which one and the same crystal call be measured different times during its rowth in the solution .
This would ful.thel render it to trace the the } ) ( sition of any given face as the crystal grows , and , if the changes are due to variations in the conditions of equilibrium crystal and tion , ably afford means for studying these conditions .
Much has been writtell concerning intel.nal structure of crystalline media , but it must not be forgotten that every point within a crystal has at some period been a point on its surface , and that our knowledge of the structure is very incomplete until the life-history of the crystal has been studied by observing the changes that take place at its surface and determining the conditions of equilibrium which obtain there .
In in the ordinary way it is very difficult to know whether the faces are jbnd until the crystal has been taken from the solution and dried , and then one cannot feel that their SUl.face has not been altered by the evaporating solution ring the process .
By neasuring crystals while still immersed solution this difficulty also would be obviated .
I feared at first that when nlnersed in a medium so a refiactive index as the saturated solution , the faces of a crystal would not yield brilliant images of the signal .
, however , is not the case ; and having convinced myself by iments that very good images call be obtained with the ordinary telescope and collimator , I had an constructed which is really nothing more than a telescope-goniometer , inverted so that the crystal is suspended in the solution .
advantage of this method is that the zone which is being measured has all its faces vertical , so that the possible distorting action of gravity is mated .
The liquid is contained in a glass vessel with plane-parallel sides , through of which the light enters normally from.the collimator , while through another the reflected light emerges normally to enter the telescope .
A grant for the purpose having been obtained from the Royal Society in 1892 , the .
was constructed according to my Messrs. TROUGHTON and SIMMS .
( 3 .
) TH GONIOMETER .
graduated of goniometer , ( Plate 13 ) , rests upon a horizontal disc which is carried by two vertical metal columns about 20 ntims .
in height ; these spring from the base plate of the instrument ; and in an aperture in one of them is fixed the collimator .
The vernier is engraved upon the rim of the fixed disc .
The telescope is fitted to the lower end of a solid counterpoised arm , which can be rotated about the goniometer axis and clamped in any desired position .
The crystal holder , which may consist of the usual centring and adjusting movements , is at the lower end of a vertical axle which passes through the centre of the goniometer disc , and may be raised or lowered by a screw with milled head , so as to bring the crystal into the horizontal plane of the telescope and collimator .
* The axle of the crystal holder is clamped by a split ring grip above the holder .
The glass trough which contains the solntion is a square vessel 6 centims. square , with plane-parallel sides mounted in a brass frame ; it stands upon a brass table , which can be raised or lowered by rack and pinion fitted below the base-plate , so as to bring the crystal into the solution or out of it .
A brass guide upon the table receives one side of the frame , For economy of , the centring movements in this instrument were not effected by the usual slides travelling between straight guides , but by two plane discs , which swing about pivots and move in contact with one anothor , as described for another instrument in ' Mineralogical Magazine , ' vol. 10 , 1887 , p. 215 .
THE VARIATION OF ANGLES OBSERVED IN CRySTALS .
and ensures that the trough shall be so placed as to have one of its faces perpendicular to the collimator tube .
The collimator being fixed , it be desired to rotate either the crystal holder or the telescope while the other parts fixed , or to rotate either of them together with the scale .
For this purpose the usual screws which bear upon the axle are dispensed with , circular metal disc , 15 centims. in diameter and 1 millim. thick , is provided , which may be clamped to the goniometer stand by one screw , to the telescope by a second screw , or to the crystal holder by a third screw .
The disc is itself in connection with the divided circle .
Thus any two parts of the instrument may be coupled together , and it may be used either as a goniometer or a spectrometer .
The telescope is clamped to the stand by a fourth screw with lever head .
Both telescope and collimator , as well as the eye-pieces , are mounted in collars provided with three screws , by which they may be fully adjusted .
The telescope is also provided with a rack and pinion , so that it may , when used as a microscope , be moyed towards or from the axis , and thus be focussed upon crystals immersed in liquids of different refractive indices .
For the measurement of crystals growing in their solutions the crystal is fixed in a small clip made of stout platinum wire or ribbon ; as the crystal grows the clip becomes enyeloped by it and affords an absolutely rigid holder .
In addition to the ordinary eye-pieces , it was found necessary at an early stage of the research to provide a special eye-piece ( fig. 2 ) for the displacement of an image in any required direction within the field of view.of the telescope .
During the growth of a crystal the asurements can only be conveniently made in a single zone , and it is inadvisable to alter the adjustment of this zone during a series of observations .
But , clearly , variations of angle may be due to the displacement of a face not only in the direction of a zone but out of it .
If a face has been adjusted so that the image ( iig .
2 ) of the square collimator signal is on the intersection of the cross wires , and if during the growth of the crystal the face alters its position slightly so that the image moves to out of the horizontal zone , it will be sufficient to measure the distance OP , and the angle which the line OP makes with the horizontal cross wire , without moving the crystal at all .
For this VOL. CCII.\mdash ; A. 3 PROFESSOR HENRy A. MIERS : AN ENQUIRY INTO pose the eye-piece is provided with an additional cross-wire MM , which can be moved across the field of view by a micrometer screw , so that the angular value of any displacement along a diameter can be measured in terms of divisions upon the head of the drum at S. The movable wire , with its micrometer adjustment , is attached to a circular frame , , which has a milled edge , and can be rotated about the centre through any required angle ; this frame is in the form of a case a circular disc which is graduated , and the angle through which it is turned is read by means of a vernier , V , engrayed upon the side of an aperture cut in the case , through which the scale is viewed .
A toothed scale at registers the number of complete turns which have imparted to the micrometer screw , each interval corresponding to one turn .
In this way the distance OP and the angle POC can be accurately measured .
is a screw by which the circle is clamped .
This eye-piece was subsequently described by Mr. TUTTON ( ' Phil. Trans vol. , 1899 , p. 463 , who added it to his crystal-grinding apparatus ; it would be a adjunct to any telescope goniometer .
For the purpose of measuring crystals immersed in the mother-liquid , one face of the glass trough having been set perpendicular to the collimator , the telescope is set perpendicular to the adjacent face of the trough by the so-called Gauss method , i.e. , by viewing the image of the cross-wires reflected at its surface by means of the light derived from a lamp placed beside the telescope and sent down the telescope by a glass plate held obliquely in front of the eye-piece .
It was found advisable for the experiments described on pp. 493-515 to fix the trough on a table which can be adjusted by means of three screws , so that a side of the trough can be accurately set perpendicular to the telescope ( see fig. 16 , Plate 13 ) .
The temperature of the solution is recorded by a thermometer immersed in it and clamped to the frame of the instrument , as shown in fig. 1 , Plate 13 .
PURPOSES FOR WHiCH THE GONIOMETER CAN BE USED .
The goniometer here described can clearly be used for any work which the ordinary goniometer is required to perform .
In can further serve several other purposes .
Crystals can be measured , as in the present research , while immersed in a concentrated solution of their own material , and the changes of form which they experience during growth may thus be studied ; deliquescent crystals , or such as are liable to ration under ordinary conditions , can be measured while immersed in oil or in some other appropriate liquid ; the action of solvents on crystals can be studied by measuring them while immersed in acids or other solvents , and this method affords the only safe means of ascertaining the angles made by the prerosion faces due to the * Another telescopic adjunct , described in the same memoir by Mr. TUTTON , as devised by myself ( p. 464 ) , namely , the swinging lens before the eye-piece , was not , as might appear from his description , lsed with this instrument , but with another goniometer .
THE VARIATION OF ANGLES OBSERVED IN CRySTALS .
action of the solvent , with the original faces , when the latter have been removed by the solvent ; the refractive indices of solid plates , prisms , or crystals can be measured by total reflection or minimum deviation by immersing them in appropriate liquids ; and the refractive index of a liquid can similarly be measured by means of a plate or prism immersed in it .
Again , the angles of isomorphous mixtures may be ascertained as the composition changes , by measuring a crystal while in a solution of different composition ; or measurements may be made of a crystal enveloping a similar crystal of a substance isomorphous with it by immersing the latter in a solution of the isomorphous material ; this would render it possible to ascertain the angles between the faces of the shell and the corresponding faces of the kernel .
Examples of some of these uses will be found in the following pages .
With the object of testing the goniometer , some observations were made the of calcite by dilute HC1 , and upon the parallel growth of sodium nitrate with calcite .
Since these are either new , or were new at the time when they were made ( 1893 ) , and were certainly made by a new method , and since they bear upon the subject of vicinal faces , a few of the results of these measurements may be here mentioned .
( A. ) Etching of HC1 .
A cleavage rhomb of calcite immersed in about 50 cub. centims. of water gave the following for two faces One drop of dilute HC1 was added and after the lapse of hours distinct pits were etched on both faces , having at first the form of linear channels parallel to the edge , and subsequently assuming a more triangular outline , and faint images begin to make their appearance close to those from and .
After 9 hours the following well-defined images were obtained in the zone [ Rr]:\mdash ; Another rhomb gave readings:\mdash ; , before etching .
.3 PROFESSOR HENRy A. MIERS : AN ENQUIRY INTO , after etching .
In another series of experiments good readings were obtained etched facets upon a cleavage rhomb immersed dilute acid ; these were inclined to the cleavnge faces at , after 2 hours , , , These images belong to facets lying in the zone and are accompanied by other nages not lying in this zone and .
to facets on the other side of the triangular etched These and other experiments indicate that well-defined facets may be developed by the action of very dilute acid , but that they are inclined at different angles to the original face , and that the inclination of the etched facets changes as the etching proceeds , the value in general increasing .
( B. ) Parallel Growth of ) A cleavage rhomb of calcite was adjusted and immersed in a concentrated solution of .
The crystals of sodium nitrate which grew upon surface of the calcite could not , of course , have all their faces parallel to those of the calcite , for the rhombohedron angle of calcite is , whereas that of sodium nitrate is It was found that the sodium nitrate always lies on the calcite in such a way that the two rhombohedra have one edge parallel , and generally also one face containing that edge in the sodium nitrate is parallel , or nearly parallel , to the corresponding face of the calcite .
This result is in accordance with that arrived at in another way by O. MUGGE , 'Neues Jahrbuch , ' 1896 ( 2 ) , p. 74 .
It is also analogous to the union of twin crystals , in which the two individuals are not always exactly in the position demanded by the law of twinning but may deviate therefrom by seyeral minutes or even by a degree or more .
In these cases the deviation takes place in such a way that the crystals retain the parallelism of at least one face or one edge , as I have shown for Proustite and Bournonite .
* 'Mineralogical Magazine , ' vol. 6 , 1884 , p. 77 vol. 8 , 1888 , p. 75 .
THE VARIATION OF ANGLES OBSERVED IN CRySTALS .
A large number of observations were next made upon potash-alum with the of discovering whether the octahedron angle of a crystal of this substance changes dul.ing the growth of the crystal , and if so , in what manner and to what extent .
The observations were carried out in the following way : a brilliant and uniform crystal having been selected , it was mounted in the platinum clip and adjusted on the instrument .
The glass trough , filled with concentrated solution , and having a few small crystals in the bottom of the vessel , was then raised until the lnounted crystal was completely immersed .
Genel.ally , the crystal would become slightly re-dissolved immediately after immersion , but it constantly watched through the telescope until its faces began to grow in a uniform manner , and yielded brilliant reflections .
It was then measured again and again at short intervals during many hours or at longer intervals during several days and nights .
The best sets of measurements were those obtained by immersing the crystal in slightly warmed and supersaturated solution and then observin it while the temperature was falling for several consecutive hours near an open window during a winter night .
With a rise of temperature on the following day the crystal would become slightly etched and re-dissolved ; but this made it possible to the changes of angle produced by incipient re-solution , and then the phenomena of growth could be watched a second time upon the same crystal during the following night .
After once the crystal has grown sufficiently to enyelope the platinum clip completely , there is no danger of its moving even to the smallest extent during any number of subsequent rvations ; it may be etched , partly re-dissolved , and allowed to grow again many times in succession without any change in its orientation .
When a crystal of potash-alum was watched in this way it was found:\mdash ; ( 1 ) That the images from the various faces continual]y altered their positions ; ( 2 ) That the angle between the adjacent faces was never the theoretical angle of the regular octahedron ; ( 3 ) That the , es usually yielded multiple images lying very close together .
Thus , in the very first crystal measured , the following readings were obtained two very good faces , A being a cube face and an octahedron face .
( theoretical angle , Next day , two images , one vertically above the other , , , , , , , beside the other , two hours later A yielded a single image , and Another crystal yielded the following from a pair of adjacent brilliant faces , and , two images , one vertically above the other , single image , , two images , one vertically above the other .
Two days later the readings had become not quite in the zone , in the zone , not quite in the zone , , , On the following day the readings were below the horizontal wire ; above , , , , , above , , , , , , , on the horizontal wire , A day later , in the morning , . .
three overlppgmages , , tree , , , , two verlapping images .
THE VARIATION OF ANGLES OBSERVED CRYSTALS .
In the afternoon On the following day two overlapping images , , , , , and so on .
This example will serve to illustrate the difficulty of adjusting and measuring one of these crystals , for images which lie on the horizontal wire at one time are subsequently replaced by images slightly out of the zone .
This difficulty necessitated the introduction of the micrometer eye-piece described above , but the observations were enough to show that each octahedron face of potash-alum is replaced by two , three , or more vicinal faces which continually change during growth .
I may now pass oyer a number of intermediate experiments , which ultimately suggested the correct interpretation of these multiple images , in order to state at once the explanation to which I was led , and will give one or two sets of experiments by which it is illustrated .
When an eye-piece of sufficient strength is employed , it is seen that many apparently simple images are in reality overlapping images lying vely close together .
These are usually three in number ; one may be brighter than the others , and one or two of them may be so faint as to be almost invisible .
Similarly , when the surface of the growing crystal is minutely examined , and with a sufficient magnification , it ' is seen that corresponding to these three images are three reflecting surfaces ; when the images are equally bright the octahedron face is seen to be replaced by three equal-sized facets of a vexy flat triakis-octahedron , as in fig. 3 ; when two are bright and the third is ahnost invisible , two of the faces are large and the third small ( fig. 4 ) ; when the octahedron face apparently yields a single image , one facet is much larger than the other two , and gives an image in the zone if the arrangement is as in fig. 5 , but an situated below the zone if the arrangement is as in fig. 6 .
This may very conveniently be done by a lens held in front of the eye-piece , THE yARIATION OF ANGLES OBSEBVED IN CRySTALS .
these are generally two in number , one flatter than the other , and replacing the upper and lower portions of the large apparent octahedron face ( fig. 7 ) .
An apparent octahedron face of alum gives , therefore , in general , not one but three images .
For the purpose of correct adjustment , and in order to distinguish easily between the various images , I have generally used a small square aperture as collimator in place of the customary Websky slit .
Two apparent faces , , of the octahedron ( fig. 8 ) , consisting in reality of the flat triakis-octahedron , will then give the reflections represented in fig. 9 .
The three images belonging to the face are not necessarily separated from each other by the same intervals as those belonging to the but the zone can be correctly adjusted by means of the images and A very perfect octahedron , after 10 immersion in solution and various changes of ture , was found on Noyember 27 , 1893 , at 10 , to yield three very well-defined images from each of four octahedron faces in the adjusted zone Calling these faces respectively , the readings and were as follows:\mdash ; Here none of the measured angles is the true octahedron angle , but if we assume that the three faces of the triakis-octahedron replacing any one VOL. CCII . .
3 PROFESSOR A. MIERS : AN ENQUIRy INTO face all belong to ] same , it is easy calculate the theorctical for the true octahedron faces ; they would be , so that The angle of the triakis-octahedron replacing each octahedron face is also easily calculated , and is for the form replacing , , , , , This example indicates that\mdash ; ( 1 ) The octahedron angle of the crystal is really the true octahedron angle ; ( 2 ) That there is a different flat triakis-octahedron replacing each face , although they only differ to a slight extent .
In this instance the images , , of each set seem to belong to the same form , for the and images are in the same vertical plane ; but this is not always the case ; the same crystal , measured on November 28 , gave the following readings from and figs. Whereas from Hence the reading for the true octahedron face is ; therefore the reading for is , and the three faces replacing are unequally inclined to , and do not belong to the same form .
If it be true that the vicinal faces , whether they belong to the same or to different forms , are always triakis-octahedron faces , then of the faces replacing an octahedron plane which should yield an image at one is always situated on the horizontal THE VARIATION OF OBSERyED IN CRYSTALS .
wire , and the other two on the lines and , which make an angle of 12 with in the of view .
This is a fact which is readily established by means of the micrometer eye-piece ; the changes of the images , for example , during the growth of the crystal can be easily followed by the movable cross-wire when it has been adjusted so as to run along the line .
This affords an even safer method of ascertaining the reading corresponding to the true octahedron face ; for after it has been found that the images from the vicinal faces travel along the lines , and , making angles of with each other , the reading for the true octahedron is known to be that corresponding to their intersection When examined this way the two adjacent octahedron faces of potash-alum give readings which differ from each other by the true octahedron angle within errors of observation ; it has thus been possible to measure with accuracy the octahedron angle of potash-alum , although the faces of the octahedron are not themselves present .
It is not necessary to multiply examples : other crystals led to the same result as that just described : namely , that the octahedron angle of potashalum is not liable to any variation , but that the octahedron faces themselves are absent and their place is supplied by vicinal planes belonging to triakis-octahedra which are liable to continual variations .
( 6 .
) THE NATURE 0F THE VICINAL ] 0N POTASH-ALUM AND AMMONIA-ALUM .
In the course of the present research many attempts made to discover some regularity in the angles or in the indices of the vicinal faces .
That they possess rational indices is rendered possible by the established fact that they lie in certain well-defined ones , but the indices are in that case very complicated .
A very welldefined set of faces , for example , on one crystal represented a triakis-octahedron inclined at 6 ' to the octahedron ; its symbol would , therefore , be ( hhi ) , where , or something approaching ( 251.251.250 ) ; but it is impossible to say whether these are the indices to be adopted , or some other numbers having nearly the same ratio .
Clearly the indices oease to be of much use when they are such PROFESSOR HENRY A. MIERS : AN ENQUIRy INTO high numbers , and the position of the faces can be equally well and conveniently represented by their angles .
In a paper on the vicinal faces of celestite Zeits .
Kryst , HINTZE suggested that their relations are better expressed by the angles that they make with each other than by the indices to which they can be referred ; and , in fact , that they do not necessarily possess rational indices .
Whether this be so or not , it will be safer for the present purpose to define the vicinal faces of the growing alum crystal by their actual angles than by their possible indices .
One reason why I am led to believe that they are really referable to rational , although not to simple , indices is the following :During the growth of the crystal , one set of vicinal faces is being continually replaced by another along certain zones ; each image of the collimator signal fades away like a dissolving view and another makes its appearance ; but the change is not gradual , neither are the surfaces curved ; one plane reflecting-surface is replaced by another plane ; and although the images may for a time be multiple and confused , sharply defined images emerge successiyely by the substitution of one image for another per saltum .
This extremely important fact is illustrated by some of the examples quoted above , and instances might be multiplied to any extent .
It is sufficient here to record the well-defined faces which have been observed during the experiments upon potash- alum and ammonia-alum .
If ( fig. 12 ) be the three images obtained from three facets belonging to the same form , then if be the position of the image which would be yielded by the ideal octahedron face , the readings actually made in the course of the observations are :\mdash ; ( 1 ) That for which coincides with the vertical cross-wire ; ( 2 ) That for which and simultaneously coincide with the vertical cross-wire .
The difference between these readings is nearly the angle between and the face which would truncate the edge ; if this face be called , the angle between and , i. e. , the inclination of the vicinal form to the true octahedron , is given by For very small angles may be taken as .
The best readings obtained on a large num.ber of crystals were the following:\mdash ; THE VARIATION OF ANGLES IN CltYSTALS .
and a number of intermediate values were obtained not quite so perfect .
The largest values observed for were on potash-alum , and on ammoniaalum , corresponding to inclinations of and , respectively , to the true octahedron face .
The following examples illustrate differences in vicinal planes replacing the four octahedron faces of one and the same crystal of potash-alum at the same time : \mdash ; These examples indicate that the angle of the vicinal faces is generally of the same order of magnitude upon the different faces at any given time , but they do not indicate any cause for the differences , or any special regularity .
In -alum I. , II .
, opposite faces seem to show nearly the same form , which differs from that on the adjacent faces ; but in -alum IV .
, and Am.-alum I. , the forms most resemble each other are on adjacent faces .
It is quite clear , at any rate , that there is no constancy of angle in the vicinal faces either of -alum or of Am.-alum .
Several attempts were made to determine whether , during the growth of a crystal , are ressively in the same direction ; whether the flat triakisoctahedron becomes more flat or more acute ; also whether the angle yaries in any way with the temperature or the barometric pressure .
The following observations on a crystal of ammonia-alum form a continuous and fairly ]engthy series:\mdash ; and are two adjacent octahedron faces and is a cube face .
and both yielded two images , , in the same vertical plane ; two images , , nearly in a horizontal plane .
, 1894.\mdash ; The crystal was adjusted for a point halfway between and and a point halfway between and all the images were very good .
At 5 was immersed in solution containing loose crystals .
Barometer 7696 millim. ; At At 10.0 yielded two images .
The crystal beeIl withdrawn , the solution then warmed to and the crystal re-immersed at 11.15 when the temperature was THE VARIATION OF ANGLES OBSERVED IN CRySTALS .
At 11.45 , the crystal was covered with etched triangles .
At 12.0 ( midmght ) , , the etched figures were gave a good at At 12.30 , the faces were quite smooth .
READINGS ( from the , \amp ; c. .
At 1.0 .
; At 1 .
10 .
; At 1.20 .
; The solution was then withdrawn .
PROFESSOR HENRy A. MIERS : AN NQUIRY INTO February 5.\mdash ; The solution having been warmed to was allowed to cool to and the crystal was immersed again at 5.30 , when it at once began to grow .
At .
; , barometer 772 millim. , THEORETICAL READINGS .
At 6.45 .
; At 7 .
10 P.M. ; At 12.30 .
; The following table sums up the results of these observations , showing the gradual THE VARLATION OF ANGLES ) IN CRYSTALS .
change in the angles made by the vicinal faces with the octahedron face they replace :\mdash ; This indicates that , as the temperature falls , the vicinal forms become in general more acuf , , their faces make larger angles with each other , but the angle does not depend entirely upon the temperature .
At , for example , is at one time , at another time something between 15 ' and 20 ; .
The same results were obtained from many other similar series of observations .
Variations of barometric pressure also seemed to be without effect upon the angles of the vicinal faces .
One suggestive feature was found to be common to many of the experiments ; it sometimes happens that an octahedron face is replaced by two different vicinal forms , one replacing the upper and the other the lower part of the face as it stands vertically in the solution ( fig. 13 ) ; in such a case it was invariably found that the triakis-octahedron on the lower part of the face is slightly flatter than that of the upper part .
For example , a crystal of ammonia-alum , immersed at midnight , , barometer 759 millims. , gay at 1.30 .
the ] owing readings for two apparently perfect octahedron faces and and for the cube face A between them .
Shortly afterwards broke up into three images , giving the readings VOL. CCII.\mdash ; A. 3 SSOR l ENQUIRy INTO At 2.30 , the readings were A A A A A o ) In addition to these images there were also three close to and , and proceeding the lower part of the face ( fig. 13 ) ; calling these and , the following measurements were obtained:\mdash ; At 2.30 S01o At 2 .
55 , .
co So that and is a flatter form than Observations of Gouy and and others have shown that solutions are more concentl.ated in .
lower than in their upper layers , and this is of course the case in the present experiments where crystals lie at the bottom of the trough .
The case just quoted , and others like it , suggest that the vicinal form is flatter where the solution is more concentrated .
Accordingly many experiments were made in which the trough was successively raised and lowered in order to ascertain the effect of immersing the same crystal to a greater or less depth in the solution .
A crystal of potash-alum gave eadings immersed deeper they became THE IATION OF OBSERVED IN CBYSTALS .
Another crystal of potash-alum , been adjusted near the surface of the solution , was immersed to a considerable depth ; the from a brilliant facet was found to moye through ( as measul .
micrometer eye-piece ) in the direction of a flattening of the form .
A third crystal gay near the surface of the liquid At a greater depth Raised to the surface again Depressed again The above experiments indicate that in crystals of potassium-or ammonium-alum the vicinal faces , which invariably replace the octahedron planes , belong to different forms of flat triakis-octahedra , and that , as the crystal grows , these continually change their inclinations ; the faces , which replace an octahedron face generally belong to the same form , and they make an angle of from 2 ' to 30 ' with ; when the vicinal facets at the top of the crystal belong to a different form from those at the bottom they generally make a larger angle than the latter with ; similarly as the crystal is immersed deeper in the solution the vicinal form becomes flatter ; during the growth of a crystal in a cooling solution the vicinal form usually becomes more acute ; but at different times quite different vicinal faces may grow at the same temperature .
( 7 .
) THE VICINAL FACES SODIUM CHLORATE , ZINC SULPHATE , AND MAGNESIUM SULPHATE .
( A. ) Sodium Chlorate .
Cubes of sodium chlorate give multiple images which , in accordance with their symmetry , are usualy two images in a horizontal plane yielded by one pair of cube .3 THE yARIATION OF ANGLES OBSERVED IN After further readjustment:\mdash ; At ; Approximate readings for true cube faces are therefore image.above c ' The crystal was re-immersed at 9.45 At 10.15 .
; ; the images from indicated a -entrant angle .
two images ; one 10 ' below , the other 14 ' above , , , , .
; ; and AB This example illustrates the difficulty of adjusting correctly a crystal which is liable to the growth of vicinal faces , and the impossibility of measuring accurately the angles between the faces of the simple forms which are replaced by them , except by tracing these changes .
In sodium chlorate the vicinal faces sometimes belong to the form , and they make angles offrom 3 to 10 minutes with the ideal cube face ; they sometimes make re-entrant angles with each other , but the difficulties of adjustment and measurement are increased by the fact that the vicinal faces often belong to various forms , and are unequally inclined to the cube faces .
The same general features hold as for alum : the vicinal forms usually become more acute as the growth proceeds ; and they are less acute for the parts of the crystal which are most deeply submerged .
A cube face of the crystal just described , for example , gave on May 27 , 1899 , two images separated by minutes when at a depth of about 1 inch , and two images separated by 12 minutes near the surface of the solution .
Zinc Sulphate and Magnesium That the same features are also exhibited by crystals belonging to other systems than the cubic , was proved by experiments on these two substances , The following are examples of ) leasul.ementsul ) hate ; isnl / NI ' On the following The value generally given for the prism angle of zinc sulphate is ; in the above crystal various prism forms are indicated , the angles of which would from to even for the best faces .
MAGNESIUM Sulphate ; prism zone ; 8.20 .
; 9 , 30 .
; above zone , , three images in vertical plane , below zone .
THE yARIATION OF ANGLEbt The value generally given the prism angle of magnesium sulphate is The above obsel.vations indicate planes to prisms whose would differ from this by more than 20 ' .
I have given these two examples from the orthol.hombic system in ' to poin out how the mean of a number of measurements may with crystals of comparatively low symmetry give a result very far from the .
In alum , although the true octahedron angle might never obseryed , mean of sufficiently large number of measurements would probably give the exact octahedron ange , for the three planes which replace the octahedron face , by the synmctry of the crystal , inclined to it in three different directions , so that some of the measured angles too large and some too small .
But in the case of a rhombic prism , each face would , in accol.dance with the symmetry of the crystal , be placed by only one and the angle , as measured , would be consistently either too or too The true position of the prism face could not be ined , as bhat of the octahedron face was in alum , by the convergence upon it of three zones of vicinal faces , by the limit to which they tend .
To determine with absolute accul.acy the true angles of such Cl.ystals it would be necessary to study the variations of their vicinal faces in the manner described ) \mdash ; but by mealls of a two-circle three-circle goniometer .
A numbel of observations were also made upon crystals of sodium arseuate of a substance belonging to the monoclinic system : but since they do not affect the general results they not described in the present paper .
( s. ) THE EFFECT 0N ALUM 0F DILUTING SOLUTION .
During the iments described in this paper there were naturally occasiolJs o11 which the Cl.ystal was not growing in it solution , but was being corl.oded by a more dilute medium , and in such cases it was sometimes possible to obtain quite good reflectiollS from the prerosion faces ; several experiments were also made in which the solution was ) osely diluted or wal.lned in order to observe the etched faces .
In the case of a bright octahedron of , the first effect of slight dilution or warmin is to develop small pits having the form of equilateral triangles inscribed in the triangle which bounds the octahedron face ; the triangles are small shallow pits and their sides must therefore be vicinal faces , which are still triakis-octahedra , like those of the growing crystal .
Sometimes the growing c1ystal has its octahedron face covered with flat triangular whose outline is parallel to that of the octahedron face ; when the solution is diluted , these have their corners rounded and become converted into hexagonal plates , then disappear and become replaced by the inscribed triangular pits .
In one case the prerosion faces ( belonging to a triakis octahedron ) on one octahedron face made an angle of with each other ; the crystal was then allowed to grow during the night , and in the morning well defined vicinal planes had made their appeal.ance , making an angle of with each other ; as the temperature rose these were replaced by prerosion faces belonging to different forms ; that on one edge being inclined at about to the true octahedron , and that on the adjacent edge at about .
This feature was observed in other etching experiments ; the prerosion faces replacing an octahedron face generally belong to different forms , whereas the vicinal planes of growth generally belong to the same form .
In one case two well defined prerosion faces belonging to the form { 221 } made their appearance upon an octahedron edge , inclined at and respectively , to the adjacent octahedron faces : the theoretical angle for { 221 } is The prerosion faces S which replace the edges of the vicinal triakis octahedron of fig. 15 belong to a different form , the icositetrahedron ; and , since the first effect of etching is to round off the edges of the crystal , it generally results in the production of such icositetrahedra ; but they do not necessarily truncate the edges , \amp ; c. Thus the crystal just mentioned yielded the following results three days later , when the corroding action of the solvent happened to be extremely slow .
, 1893.\mdash ; At 10 , each octahedron face yielded three very perfect images ; the actual readings are given on p. 473 .
A little later , the form replacing was measured and was found to be a triakis-octahedron inclined at ( previously determined as 5 At 10.45 the edges of this vicinal form were replaced by facets of an icositetrahedron S inclined at almost exactly the same angle to ; and the same thing happened on the other faces of the crystal .
The angle between S and the true octahedron face was measured as on on ; on .
The triangular etched pits did not begin to make their appearance until 3 On the following day the same events recurred : the vicinal faces of growth eplacing o belonged to a triakis-octahedron inclined to at , and at 11.15 the edges of this form were replaced by prerosion faces belonging to an THE VARIATION OF OBSERVED IN CRYSTALS .
icositetrahedron inclined to at .
It is , no doubt , the development of such faces which conyerts the triangular plates of growth into hexagonal plates , as described above .
These and other instances indicate that the first effect of slow solution is to produce prerosion faces on the edges of the crystal ; on the octahedron edge these belong to triakis-octahedra ; on the edges of the vicinal faces they belong to icositetrahedra .
The latter are inclined to the octahedron at about the same angle as the triakis-octahedron , they certainly do not truncate its edges .
The next effect of the process of solution is to produce etched triangular pits bounded by triakis-octahedra .
( 9 .
) CAUSES 0F THE VABIATIONS .
With all the substances examined , as with alum , vicinal faces are the rule ; they change their position as the crystal grows ; they to deviate more widely from simple forms , at any rate during the initial growth ; reflections move or oscillate generally in well defined zones , and in accordance with the symmetry of the crystal ; their movement is not gradual but per saltum ; their deviation from the simple forms which they replace generally amounts to from 3 to 20 minutes .
In making the observations great care must be exercised lest deviations due to refraction through layers or currents of more or less concentrated liquid be mistaken for those due to the change of inclinations in the vicinal faces .
Such currents are always streaming round a growing crystal , and do produce visible displacements of the images ; but these displacements are rather of the nature of irregular oscillations , and may be detected with a little practice .
When beginning the study of alum , I to eliminate action of these currents with the object of ascertaining whether the production of the vicinal faces might not actually due to them .
By means of a suitable c]ockwork arrangement a wooden yane with four arms was kept revolving in a beaker full of concentrated alum solution during several hours , so as to keep the liquid continually stirred .
Small alum crystals were allowed to crystallise on the yane , and were examined at intervals .
I came to the conclusion that the crystals were equally perfect , and possessed vicinal faces of the same sort , whether they crystallised in a still solution or in one which was kept continually stirred .
In the same way the stirrin of the liquid in the goniometer trough seemed to produce no appreciable effect upon the nature of the vicinal faces on a crystal of alum which had been growing in the still solution ; introduction of a Qtirrer into the trough did not affect the smoothness or the angles of the faces so far as could be observed ; the are , however , difficult to carry out , and it is of course impossible to make accurate measurements while the liquid is in violent motion , although the crystal itself is not disturbed .
* * Other and more precise experiments of this nature are described below on p. 518 , VOL. CCII.\mdash ; A. 3 PROFESSOR HENRY A. MIERS : AN ENQUIRy INTO Among other possible causes to which the production of particular vicinal forms might be ascribed are\mdash ; ( l ) the temperature at which growth is taking place , ( 2 ) the rate at which the temperature is changing .
It became important to examine next how far these two factors might contribute to the results observed .
From the example quoted on p. 479 and from other similar examples we may endeavour to trace a connection between the angle of the vicinal form in alum and the rate of fall of temperature .
represents the theoretical octahedron face , the vicinal plane .
From the above observations it is very difficult to draw any definite general conclusions , since the differences between the various angles in same crystal are greater than the differences between the mean values in different crystals .
This may indicate that the variations are mainly due to local causes .
On the whole , the made by the vicinal forms with the octahedron is perhaps slightly greater when the temperature is falling rapidly ; and for stationary temperatures it is perhaps slightly .
greater for low than for high temperatures ; but the differences are very small , if , indeed , they really exist , and in many cases quite different vicinal faces are developed at the same temperature , so that in any case the angle of the vicinal form can hardly be a direct result of the particular temperatul.e at which it was Having found , however , that there seemed to be a general relationship between the acuteness of the vicinal faces and the depth at which the crystal was growing in the solution , I was naturally led next to enquire whether one determining factor is not the concentration of the solution ; it is certain that the concentration varies slightly with the depth , especially in solutions which contain loose crystals at the bottom of the trough , and the concentration seemed to be one of the variable factors capable of accounting for the changes of angle .
PART II .
THE CONCENTRATION 0F THE SOLUTION 1N CONTACT W1TH A GROWING CRYSTAL .
( A. ) Prelimina ry It was pointed out above that in studying the growth of a crystal we are dealing with the conditions of equilibrium between the growing crystal and the liquid in contact with it , and it is therefore a matter of primary importance to know all that can be ascertained about the latter .
Very little seems to be known at present about the nature of this liquid ; that it is supersaturated is a reasonable supposition , but to what extent is not certain .
The solution is being constantly impoverished by the growth of the crystal , although it is .
constantly enriched by the influx of more saturated liquid .
Among the possible factors which determine the different vicinal faces succeeding each other during the growth of the crystal , it appeared from the experiments described above that differences in the degree of concentration of the liquid may conceivably play an important part ; but in order to test this possibility it would be necessary to ascertain the composition of the liquid , not merely in the neighbourhood of the crystal , but actually in contact with it .
Two methods of attacking the problem suggest themselves ; although one cannot hope to analyse the layer of liquid THE VARIATION OBSERVED IN 49.3 in contact with the crystal , it may be possible to measure either its specific gravity or its refractiye index , and from these to deterlnine the degree of concentration .
Now goniometer used in the earlier part of this research lends itself ably to the measurement of the relractive indices of solutiolu by the method of total reflection .
If a glass prism be adjusted on the crystal holder , its .
and refractiye index can be measured ; it may then be immersed in the solution , and if we substitute for the square previously used a cylindrical trough ( fig. 16 , Plate 13 with a plane glass front set perpendicular to the telescope , it will not be difficult to measure the angle of total internal reflection in the prism , and so the index of refraction of the liquid .
Furthel , by using a sufficiently small prism , the index of the liquid may be ascertained at any desired point , and by suspending a crystal in the solution and moving it gradually towards the face of the prism , it may be possible to explore the liquid and to trace the change in refractive index as we approach the growing crystal .
A few obsel'vations were made in this way with a small glass prism of about and a crystal of alum immersed in concentrated solution in the square trough of fig. 1 .
A face of the crystal and the reflecting surface of the prism were accurately adjusted parallel to each other in the following way .
The crystal was fixed for purpose on a wire rod curved over the edge of the trough and mounted in a crystal holder provided with the ordinary tangent and traversing screws for centring and adjusting ; the holder was clamped to a heavy stand placed beside the goniometer .
The crystal was first adjusted until it reflected the of the collimator slit upon the cross-wires of the telescope ; it was then parallel to the axis of the goniometer .
The crystal having been withdrawn to a short distance by the traversing screw , the prism was next adjusted so that its hypothenuse or totally reflecting surface also reflected the collimator signal on to the cross-wires .
The prism being then turned through had its reflecting surface parallel to and facing the crystal plane , and the crystal could be brought as close to it as is desired by means of the traversing screw .
The telescope was next set at right angles to the second prism face , and then moved from this position through an angle equal to the angle of emergence for the saturated solution ; the front of the trough being also moved so as to remain perpendicular to the telescope .
Finally , therefore , the prism and telescope are in the approximate position for measuring the index of the liquid , and the crystal , having a face parallel to the reflecting surface of the prism , can be brought into absolute plane contact with it .
The trough is a vessel with truncated ; it is held in position by three clips , two of which press the plate of glass which constitutes the front against the edges of the vessel , and prevent kage .
The trongh is djusted upon its table by three levclling this table is fixed in any desired position upon the table of fig. 1 by means of two catches at the bnck and a clamping screw in front .
PROFESSOR HENRY A. MIERS : AN ENQUIRY INTO It was found that a perfectly definite reading being obtained with sodium light for total reflection\mdash ; corresponding to the index of the saturated or slightly supersaturated solution\mdash ; this reading was only slightly aflected when the crystal was moved up and even brought into actual contact with the prism face .
The index does actually increase to a certain small extent as the crystal is brought nearer to the prism , and this is shown by Nos. 1-4 in Table II .
, p. 500 .
This indicates that the index of refraction only varies to a very slight extent , and therefore that the solution is only slightly supersaturated almost up to contact with the growing crystal .
It is clear , however , that this experiment does not really determine the index of the liquid in absolute contact with the crystal , whereas the ultimate layer is precisely that of which we require to know the composition , even if it be only a layer of in ppreciable thickness .
Is it not possible , however , to make use of the natural faces of the growing crystal as our totally reflecting prism ?
This , if feasible , appears to be in fact the only method of determining the refractive index of this ultimate layer of the liquid .
Preliminary experiments , in which the square glass trough and an octahedron of potash-alum were employed , presented considerable difficulties , but ultimately yielded a definite and clearly visible line of demarcation in the field of view , and showed that the method is a possible one .
Let OAO'B ( fig. 17 ) be the outline of four faces of an octahedron of alum adjusted upon the goniometer , and growing in the solution ; being the telescope and the collimator .
The angle between the telescope and the collimator is first accurately measured by means of a glass plate mounted on the crystal holder , being twice the angle between the position in which the glass plate reflects the collimator signal on to the cross wires , and that in which the plate is perpendicular to the telescope as determined by the Gauss method .
The front HK of the glass trough , HKNM , is then adjusted perpendicular to the telescope by the Gauss method , and the angle AOB is measured .
A movable sodium flame at is next adjusted by trial , and the crystal rotated , until the edge of the dark band indicating total internal reflection at OA is visible in the field of view and can be adjusted upon the vertical cross-wire .
Let the reading for this purpose be ; then , if be the reading for which the image of the collimator signal is reflected upon the vertical wire from OB , and the reading for which OB is perpendicular to , then is the angle of THE yARIATION OF ANGLES OBSERVED IN YSTALS .
emergence of the reflected light ; i.e. , the angle between the rays entering the telescope from the crystal and the normal to the face OB .
If , then , A be the angle AOB , and be the refractive index of the crystal , the refractive index , , of the liquid is given by the formula ; where tall A preliminary ent was made on November 14 , 1900 , with a good Cl.ystal of potash-alum growing at a temperature of The refractive index of the alum crystal for sodium light was measured in two ways : ( 1 ) by minimum deviation , using octahedron as refracting prism ; ( 2 ) by total internal reflection , makin use of the natural faces of the .
and was found to be by minimum deviation , by total refJection .
The values previously found by STEFAN , , MULHEIMS , , FOCK , F. KOHLRAUSCH vary between and The refractive index of a solution saturated at was also mined in two ways : ( 1 ) by minimum deviation , using a hollow glass prism filled with the solution ; ( 2 ) by total reflection , using a glass prism of refractive index immersed in the solution contained in the square trough , and pursuing the process ibed above .
It was found to be at by total reflection , at by minimum deviation ( the latter corresponding to at ) .
With a prism of formed by two octahedron faces , and with for the index of the crystal , and for the of the liquid , the angle of emergence should be Of course one difficulty introduced by the use of the alum crystal itself is that the angle is not exactly that of the regular octahedron , and is liable to small val'iations .
The faces , , of the crystal used in this experiment did not give multiple images , and were inclined to one another at the angle .
For this prism the angle of emergence should be The preliminary experiment gave a smaller angle of emergence , indicating a higher refractive index for the liquid , and showing that the solutio]l in contact with the gro is , in fact , supersaturated .
Before proceeding to determine the exact refractive index of the liquid , it now necessary to ascertain ( 1 ) how the refractive index of strong solutions aries with their concentration ; ( 2 ) whether there is any sudden change in the index in passing from the saturated to a supersaturated solution ; and ( 3 ) how the refractive index of supersaturated solutions varies with their strength .
So far as I am aware , no previous observations have been made upon the refractive indices of supersaturated , or even of saturated , solutions of the substances inyestigated in this paper .
( B. ) The tdices of Solutions .
The observations were made in two ways : ( 1 ) By direct measurement of the minimum deviation in a hollow prism containing the solution ; ( 2 ) By total reflection , by means of a glass prism of a known index immersed in the solution , using the new goniometer as described on p. 493 .
With the minimum deviation method a supersaturated solution containing a known weight of salt to a known weight of water was taken , and a sample of about 6 cub. centims. introduced into the prism ; the solution was then diluted with a measured volume of water and another sample examined , and so on .
The temperature was measured by a thermometer inserted in the prism .
With the total reflection method it is easier than with the minimum deviation method to carry out a long series of observations upon solutions of slightly varying strength ; for it is only necessary to add known volumes of water successively to a known solution .
For example , 50 cub. centims. of a supersaturated solution of known strength having been placed in the glass trough , its refractive index was determined by the totally reflecting rism ; 1 cub. centim. of water was then added and the index re-determined ; this process was repeated until the trough was nearly full ; a measured volume of the liquid was then removed from the trough , and the process was repeated and carried on until the solution was quite dilute .
The liquid was , of course , kept constantly stirred during these experiments .
As regards the accuracy of the two methods , several examples will be found below of.determinations of the same liquid by both ; these always agree in the third place of decimals , and sometimes in the fourth .
As regards sensibility , an error of 1 minute in the reading for total reflection corresponds to about 2 in the fourth place of decimals , or to an error of about 2 minutes in the reading for minimum deviation .
Although the readings for total reflection cannot be made so accurately as those for minimum deviation , they could generally be relied upon to within 2 minutes , and sometimes to 1 minute .
Preliminary observations were first made upon solutions of alum and of tartaric acid , chiefly with the object of plotting the results upon a curve in order to ascertain whether there is any sudden change in the curvature , especially on passing from a THE TION OF RVED IN CRYSTALS .
supersaturated to a saturated solution .
In these experiments the constitution of the original saturated solution was not known , but that of the diluted solutio1l at the close of the series was ascertained , and the former was approximately calculated from it .
( 1 .
) of Solutions of -almn .
The refractive index of a saturated solution of potash-alum was approximately determined to be at by total reflection , at by minimum deviation .
The following numbers represent the rvations made by the total reflection method , with a prism of angle and index , upon 50 cub. centims. of a saturated solution of alum diluted down by successive additions of 1 cub. centim. of water ( up to and then of 5 cub. centims. ( up to and finally of 25 cub. centims. VOL. CCII.\mdash ; A. SSOR HENRY A. IsIIERS : AN NQUIRY INTO the total reflection method be relied upon , and that there is no sudden break the curve .
In this particular series there is a considerable dis]ocation of the curve between Nos. 10 and due partly to the fact that the corner of the rectangular trough used in this experiment interfered with the illumination , and partly that , for fear of , the liquid to crystallise , it was not efficiently stirred until this point was reached .
For this reason the series is not plotted on Diagram I. Table II .
represents a series of isolated observations made by the method of total reflection upon saturated solutions at the ordinary temperature .
In Nos. 1 to 4 , which relate to the same solution , the temperature was not recorded ; but the effect of approximation to crystals in the solution is here traced ; No. 1 is a saturated solution ; in No. 2 a few small crystals have been placed in the trough ; in No. 3 mounted crystal is brought near to the prism ; in No. 4 it is brought almost into contact with the prism .
All these are solutions saturated , or very slightly supersaturated , at the ordinary .
but the observations are not represented on Diagram I. because the exact composition of the several solutions is not certain .
Table III .
represents a series of observations made upon a supersaturated solution of known gradually diluted down by addition of water , and determined by means of a hollow prism holding about 6 cub. centims. of liquid .
These are denoted by small circles on Diagram I. , and serve to establish approximately a curve which shows the relation between composition and refractive index in supersaturated and strong solutions of potash alum .
Table IV .
represents a few isolated observations made by the hollow prism upon solutions of known strength ; they are denoted by small crosses on Diagram I. , and serve to check the curve represented in Table III .
Strongly sulJel.saturated solutions of unknown constitution were next examined by the teflection method with the prism of angle and index 1.51704 .
In I. , 50 .
centims. of a supersaturated solution were placed in the diluted by successive additions of 2 cub. centims. of water .
The results in V. , and are represented by dots in Diagram II .
The strength of these solutions is obtained by assuming that No. 6 has the same constitution as No. 1 in Table VIII .
, which possesses practically the same refractiye index , i.c. , that 50 cexltims .
weigh grammes and contain grammes of the substance .
50 cut ) .
centims. of the solution were diluted with 25 cub. centims. of water ; and 50 cub. centims. of this new solution ( 2 ) were then examined in the trough and diluted by successive additions of 5 cub. centims. of water .
The results are given in Table VI .
, and are represented by dots in Diagram II .
: this set should , of course , be continuous with the preceding ; they are connected by a dotted line on the diagram .
In Series II .
, 50 cub. centims. of a supersaturated solution were diluted by successive additions of 5 cub. centims. of water .
The results are given in Table VII .
, and are represented by crosses in Diagram II .
The strength of these solutions is obtained by assuming that No. 3 has the same constitution as No. 5 in Table V. , the two having nearly the same refractiye index .
PROFESSOR HENRy A. MIERS : AN ENQUIRy INTO Table VIII .
represents a series of observations made upon a solution of known strength gradually uted down by addition of water and determined by means of the hollow prism , as in the case of alum .
These are represented .
by ] circles on .
and serve , as before , to establish approximately a curve showing the relation between composition and refractive index in supersaturated and strong solutions of tal.taric acid .
ersaturated and strong solutions were next examined by the total reflection method , as in the preceding examples .
In Series I. , 20 cub. centims. of a solution containing grammes of sodium chlorate were placed in the trough and examined with a totally reflecting prism of , and index , and diluted down by successive additions of 2 cub. centims. of water .
The results are given in Table IX .
, and are represented by dot Diagram 1II .
In Series II .
, 10 cub. centims. of a supersaturated solution containing gramme 508- ROFESSOR Ry A. MIERS : In Series II .
, 20 cub. centims. of a saturated solution were examined and diluted down by successive additions of 2 , 4 , 6 , or 8 cub. centims. of water .
This series of observations was made in a single day .
The results are given in Table XIII .
and are represented by dots in Diagram IV .
In Series III .
, 10 .
centims. of a supersaturated solution , weighing grammes and containing grammes of substance , were placed in the trough and diluted by successive additions of 2 cub. centims. of water up to No. 12 ; warm water being used up to No. 5 ; water at the ordinary temperature from No. 6 onwards ; Nos. 13 and 14 were made after the addition of 4 cub. centims. each .
The results are given in Table XIV .
and are represented by dots in Diagram IV .
The irregularity in the series of crosses is fully accounted for by the break in the continuity of the observations .
HENRy QUIRY INTO to no sudden in the curvature of the curve which expresses the betw refiactive index and constitution .
They must not be arded as more than a first atteml ) to establish this fact and to determine the curve for four substances : in the case of the supersaturated solutions , the stirring was not always satisfactory , on account of the danger of making the solution crystallise ; and in the same series of obseryations the constitution was only deduced by comparison with another series .
However , the very fair coincidence between the determinations made quite independently by the two methods giyes some confidence in the experiments next to be described , in which a growing crystal itself is employed as the totally reflecting prism .
( C. ) efractic e Index Concentration of the Solution in Contact with Growing Crystals of Alum , Sodium Sodium Nitrate .
As explained above , the growing crystal itself may in some instances be used as the totally reflecting prism , and it will then give the refractive index of the solution in contact with the crystal by the formula on p. 495 .
Some difficulty will , of course , be experienced by reason of the vicinal faces ; in the first place , the angle of the prism have to be measured during , or immediately after , each observation ; this can be done without withdrawing the crystal from the solution if the square trough be employed ; in the second place , and this is a more serious difficulty , the prism faces are no longer single plane surfaces , but sets of vicinal faces , and it will be necessary to select such as are very flat , or to make use of the growing crystal at the precise moment when two large vicinal faces can be employed as a prism and the rest are sufficiently small to be ignored .
In spite of these difficulties , observations have been made upon the two cubic substances , alum and sodium chlorate , and upon sodium nitrate , which serve to determine approximately the refractive index of he liquid , and therefore , by the help of the preceding diagrams , its constitution .
Since the light , on emerging from the ystal , trayerses a liquid of varying density , and since the crystal face through which it emerges is not parallel to the side AB of the trough nearest to the telescope , the formula on p. 495 is not rigidly applicable ( see fig. 18 ) .
If be the index of the crystal , the unknown index of the solution in contact with it , and the index of a saturated solution , which is known from the preceding VARIATION OF ANGLES RVED IN CRYSTALS .
experiments , then the light sfter emergence traverses a liquid whose index from to ; and , if it be assumed that it changes , then the index caIl be deduced from the angle of emergence and the angle of the crystal by the formula Under the conditions described below , this formula is practically identical with that given on p. 495 .
( 1 .
) With alum either the acute or the obtuse angle of the octahedron may be employed , and , thel.efore , a theoretically pel .
fact crystal would admit of foul determinations without readjustment , using in turn the four lron faces as the totally reflecting surface .
In general it was not possible to carry out more than one determination on each crystal in consequence either of vicinal , or the difficulty of illumination owing to the shape of the crystal , or the imperfections of the faces .
A very perfect crystal giving single reflections from two faces and was adjusted , and immersed in concentrated ] ution one evenin and allowed to grow during the night ; on the following day no reading could be obtained , but on the second day a reading was obtained for total reflection at , the light emerging through ( fig. 19 ) .
Assuming that the angle is , and calculating from the reading for which was perpendicular to the telescope ( itself being a somewhat unsatisfactory face ) , the angle of , was found to be .
Taking the value for the index of the alum , this leads to the value for the index of the liquid in contact with and , therefore , dicates that the liquid is only slightly supersaturated .
If index of the liquid had been , the value found for the solution , the angle of emergence would be , a difference lying .
beyond the errors of observation .
If the angle of the prism , as deduced from the measured angle instead of the theoretical angle , the angle of emergence would been , indicating that the difference observed cannot be attributed to a small error in the angle of the crystal .
The crystal was re-examined in the evening , and , immediately after the for total reflection had been taken , the crystal soft tong)ue of blotting paper and re-measured .
The faces CD yielded ( in to other images ) two bright images corresponding to faces inclined to each other at and the angle of emergence through was .
This gives the index as A third determination was made with the same crystal on the following day , using the acute angle of the octahedron as prism , i.e. , total reflection taking place at and emergence through C. This gave and the angle of emergence , leading to the value for the index of the liquid ; in this case , however , the crystal was probably not growing , but was in contact with a diluted liquid , for the line of total reflection was not well defined , the addition of a little water did not affect the reading , ond etched triangles made their appearance on the faces .
The value is simply the index of an almost saturated solution .
Another crystal was mounted and examined in the same way with the obtuse angle for prism ; this gave The same crystal , examined on the following day at , gave But the faces now yielded triple images , so that this determination is possibly not so reliable as the preceding one .
The crystal was then withdrawn and the glass prism inserted in place of it ; this gave The above observations are sufficient , I think , to establish conclusively that the liquid in contact with a growing crystal of alum is slightly supersaturated , and that its index differs by about 0007 from that of a saturated solution at the same temperature .
Other observations led to the same result .
The mean of the best determinations for the solution in contact with a growing crystal at is The best ation for the saturated solution at gave these results by the curve of Diagram , we arrive at the conclusion that at C. : \mdash ; THE VARIATION OF RVED IN CRySTALS .
( 1 ) A saturated solution of potash-alum contains about 8 .
ammes of the substance in 100 grammes of solution ; ( 2 ) The liquid in contact with a growing crystal contains about grammes of the substance in 100 grammes of the solution .
Specific of Solution in mith Crystal of The refractiye index of the liquid in contact with a growing crystal having been determlned , it seemed desirable to determine also its specific gravity .
A solution was therefore made containing grammes of -alum per 100 grammes of solution , and the following obseryations were made : Temperature .
Specific grayity .
referred to water at C. This agrees very closely with the value obtained by constructing a curve from the values found by GERLACH and quoted by LANDOLT and BoRNSTEIN , 'PhysikalischChemische Tabellen , ' 1894 , p. 203 .
The refractiye index of this solution was determined by the method of total reflection with a glass prism of , and refractive index , immersed in the liquid .
The value obtained At . . . . .
The index , as determined from the curve of Diagram I. , would be at C. ( 2 .
) Sodium Chlorate .
Sodium chlorate has a refractiye index of ] .
for sodium light , according to KOHLRAUSCH , and we have already found that the index of a saturated solution of the substance at is about ; with these values , and using cube faces for the surfaces of the reflection and emergence , i. e. , using a prism of angle , the angle of total reflection should be , and the angle of emergence A fairly good cube of sodium chlorate growing in its solution yielded readings for total reflection from each of the four cube faces in one zone ; the best of these corresponded to the angle of emergence Assuming the to be , and taking the value for the index of the crystal , we get for the index of the liquid in contact with the crystal .
A second crystal gave , and , therefore , A third crystal gave the following very good readings at , the observations being made late at night when the crystal was growing steadily and VOL. CCII . .
3 PROFESSOR HENRy A. MIERS : AN ENQUIRY INTO Readings for the images of the collimator signal reflected from the four cube faces As explained on p. 484 , the faces yield each two images lying in a horizontal plane , and the faces two images in a vertical plane .
The readings for total reflection were:\mdash ; light reflected at emerging through A. , , , , .
The angle between telescope and collimator was On the following morning almost precisely the same readings were obtained .
There is clearly here a difficulty in deciding what to take as the value of he angle Confining our attention to the case in which total reflection takes place at the face , and taking in succession the maximum value , the minimum value , and the mean value AB , the three results at Finally , then , we have for the refractive index of the liquid in contact with a growing crystal of sodium chlorate at about .
Interpreting these results by the curve of Diagram III .
we arrive at the conclusion that at ( 1 ) A saturated solution of sodium chlorate contains about grammes of the substance in 100 grammes of solution ; ( 2 ) The liquid in contact with a growing crystal contains about grammes of the substance in 100 grammes of solution .
THE VABIATION OF ANGLES OBSERyED IN CRySTALS .
( 3 .
) Sodium The preceding observations relate to isotropic substances ; it appeared advisable to apply the same method to some substance not belonging to the cubic system and possessing strong double refraction .
Sodium nitrate lends itself admirably to this purpose and both cleaves and crystallises readily in the form of rhombohedra of .
The refractive index of a saturated solution of nitrate was determined above to be about at ( mean of .
With a prism of this would give for a cleavage rhomb of sodium nitrate in contact with a saturated solution an angle of total reflection of and an angle of emergence of , taking , and using only the ordinary index ( see fig. 20 ) .
A cleayage crystal having an angle of between two very good faces was immersed in saturated solution and gave an angle of emergence of .
This leads to the value for the refractive index of the liquid .
Another cleavage fragment gave at A third cleavage fragment immersed in a saturated solution kept at about the same temperature by surrounding the trough with kers of ice gave A fourth , for which the solution was cooled down to by the same method , gave A fifth gave Finally , then , we have for the refractive index of the liquid in contact with a growing crystal of sodium nitrate mean of and , at Interpreting these results by the curve of Diagram IV .
we arrive at the conclusion that at :\mdash ; ( 1 ) A saturated solution of sodium nitrate contains about grammes of the substance in 100 grammes of the solution ; ( 2 ) The liquid in contact with a owing crystal contains a.bont rammeRof the substance in 100 grammes of the solution : It must be remarked that owing to the use of cleayage fragments in these experiments it is not absolutely certain in every case that the crystal was actually growing , and it is possible that the higher values are the most correct .
It will be noticed that the refractive index of a saturated solution of sodium nitrate is greater than the extraordinary index , , of the crystallised substance , and less than the ordinary index , , so that only the ordinary index comes into operation in these experiments , and with regard to the index of the crystal no account be taken of the direction in which the light is travelling in the crystal .
With all the three substances examined the result is the same : the liquid in contaot with a growi1lg crystal is slightly , and only slightly , supersaturated , PART III .
THE ACTION 0F THE CONCENTBATION STREAMS ( CRYSTALLISATION 1N MOTION ) .
of the Foregoing Results .
A Possible Cause for the of Angle in other In the preceding pages we have seen that , in the case of alum , the apparent variation of the ootahedron certainly due to the production of vicinal faces , and that the same cause is probably sufficient to account for the variations of angles in the forms of other substances , The variations are small , but they lead to the following important results : \mdash ; The faces which actually occur upon crystal are , in general , not those with simple rational indices , but are faces .
It is true that these probably possess rational udioes , but such as can only be expressed by high numbers .
To what are these faces to be attributed ?
There is little in orystallographio literature which throws light upon the subject , and , so far as I am aware , absolutely no experimental work .
Only in the recent remarkable and original investigations of G. WULFF , upon the velocity of crystal growth , is a definite explanation of the vicinal faces propose WULFF finds that the velocity of outward growth of a crystal is different for different faces , and he supposes that it is probably inversely proportional to their reticular density ; the actual velocity of growth is , however , largely influenced by the intensity and direction of the concentration streams which flow upwards round the crystal .
Experiments upon these led him to the conclusion that the crystals ' ' Sir ]irage Geschwindigkeit des Wachsthums llnd de-r Auflosung der chen , \ldquo ; ' Zeitsch . .
Kryst 1901 , yol .
34 , p. 449 .
THE yARIATION ANGLES OBSERVED IN most regularly when the concentration streams are weakest .
In order to eliminate the effect of the concentration streams , his crystals were made to grow in rotating vessels , and very uniform results were obtained .
WULFF points out that those faces which grow outwards most slowly increase in area at the expense of those which grow most rapidly , so that the latter are soon eliminated ; the actual forms occurring on a crystal are , therefore , those whose rate of growth is least , and these are the faces of greatest reticular density .
The substance used by WULFF in his experiments was ferrous-ammonium sulphate .
In an investigation by Z. the experiments were extended to iron- ammonium-alum , and led to the result that crystals of this substance grow in an aqueous solution , with a supersaturation of about 2 per cent. , and at a temperature of C. , twice as fast in a direction perpendicular to the dodecahedron faces as perpendicular to the cube faces , and from three to four times more rapidly perpendicular to the cube than to the octahedron .
According to , therefore , the usual planes on a crystal must be those of high reticular density , and the vicinal planes must be exceptional and due to some disturbing cause , which he describes as follows:\mdash ; " " Die Concentrationsstromungen verursachen das Erscheinen der chen .
Bei der Krystallisation orientiren sich die Molekeln auf den Flachen des Krystalles ganz gleichformig durch den Einfluss der Richtkraft der Krystallisation .
" " Jedoch mussen sich diese Krafte unbedingt mit den ften , durch welche die Molekeln von den Concelltrabionsstromungen fortgezogen werden , zu einer Kraft zusammensetzen .
Die sich ablagernden Molekeln orientiren sich nach den Richtungen , die den resultirenden ften ]ltsprechen , ulld am End lt man Flachentheile , welche um einen mehr oder weniger kleinen Winkel aus normalen Lage gebracht worden sind .
Dieses sind die Vicinalfliichen .
Es ist klar , dass die Lage solcher chen volume folgenden Factoren Von der Intensitat der Stromungen ; ( 2 ) von der Richtung der Stromung in Bezug auf die normale Lage der Flache , und ( 3 ) von der der Flache , welche selbstverstandlich in einem mehr oder weniger directen Zusammenhange mit der des Krystalles steht .
Spaterhin werden .
sehen , dass die Energy des Wachsthums der Flache , bis auf einen gewissen Factor genau , der sich in nicht weiten Grenzen andert , umgekehrt proportional sir ren Dichte der iche des Krystalles steht .
Bei sehr geringer reticularer Dichte , wie sie sich fiir die ergiebt , wenm man ibre Symbol nach den beobachteten Winkeln berechnet , wird diese Energy so gross , dass eine solche iche verwachst und Deswegen darf man die Vicinalflachen nicht als chen mit complicirten Symbolen betrachten , wie es sehr viele Krystallographen zu thun pflegen .
darf man die Vicinalflachen als Flachen mit einfachen Symbolen ' Studien uber relative Wachsthumsgeschwindigkeit der chen , \ldquo ; , p. 531 .
PROFESSOR HENRy A. MIERS : AN ENQUIRy INTO auffassen , welche aber , nach der Terminologie JEROFEJEW'S , zu verschiedenen ' zusamiuften ' Individuen gehoren , und zwar deswegen , weil die Winkel dieser ichen nicht genau den idealen Grossen , welche die Symmetry erfordert , entsprechen .
Diese Ansicht ist das Resultat eines falschen Begriffes vom Krystalle als von einem Inlividuum , das sich unbedingt durch einen bestimmten Anfangspunkt des Wachsthums auszeichnet .
Wir konnen den Krystall nur dann als aus mehreren Individuen bestehend betrachten , wenm man in ihm mehrere solche Punkte nachweisen kann .
Andererseits erhalt man den Begriff ' Zusammenhaufungen der Individuen ' in Folge einer einseitigen Vorstellung vom , wenn man ihn nur als geometrischen rper betrachtet .
Beide Ansichten uber die Vicinalflachen haben nur dadurch entstehen konnen , dass man den Krystall an und fiir sich , unabhangig von dem Medium , in welchem er sich bildet , betrachtete Now the question whether the vicinal faces are due to the concentration streams is one in which direct appeal can be made to experiment .
I have mentioned above the experiments which I made in 1894 with the object of eliminating possible differences of concentration in the mother-liquor .
Small crystals of potassium-alum were suspended from the arms of a wooden vane , which was kept rotating in a concentrated solution of the salt for several by means of clockwork .
No effect whatever seemed to be produced upon the crystals either as regards the presence or absence of vicinal faces or the brilliancy and smoothness of the crystals .
Recently I have made similar , but more cise , experiments in order to ascertain whether the stillness or violent motion of the liquid have any effect in altering the angles of the vicinal forms .
The experiments were made in two ways : a crystal , having been mounted in the platinum clip , was allowed to grow until well developed , and was measured in the solution ; without being withdrawn from the solution , it was then attached to a vertical spindle and kept rotating for several hours by means of a small and was then measured again .
In other the trough of liquid was placed upon a turn-table and set rotating , the crystal remaining fixed within it .
Experiments upon -alum in Agitated An octahedron yielded three well defined vicinal planes upon each of two octahedron faces A and : the angle between one facet and the edge of the other facets , , was measured as in fig. 10 .
: ; : At 4 .
the crystal was set rotating , the temperature being C. , the crystal was stopped and measured , and gave : ; : ; but these were now accolnpanied by other and less distinct images .
THE VARlATiON ANGLES OBSERVED IN CRYSTALS .
The crystal was re-immersed at 7.45 , and allowed to rotate during the whole night .
Next morning at 9 .
SO , it was measured again and gave upon another octahedron face these readings being again very good .
The same crystal was re-immersed at 3.30 , and the trough was set rotating .
At 9 .
, it gay the readings being very good .
The corresponding readings for the other vicinal forms were : Before tion .
After rotntion .
On . . .
, , . . .
\ldquo ; . . . . .
As a result of these experiments it may be asserted that , in alum at any rate , the vicinal faces are neither produced nor appreciably affected by the concentration streams in the solution .
The vicinal faces may , howeyer , vary with average concentration of the solution in their neighbourhood , or\mdash ; what may amount to the same thing\mdash ; with their rate of growth .
In fact , it is difficult to think of any other nearly constant , but slightly variable , condition to which they may reasonably be attributed .
In any case , the fact to which I wish to draw attention is that the faces which actually occur upon a crystal are not those with simple indices and great reticular density , but those with complex indices and low reticular density .
The faces of alum whose angles were measured by BRAUNS , and whose rate of growth was measured by WEYBERG , were , therefore , not octahedron faces at all , but vicinal faces , and in the arguments relating to them we are not at liberty to assume that they are faces of high reticular density .
Whatever structures may be necessary to account for other features of crystals , there is little doubt that we are justified in regardin their faces as the of a space-lattice .
Now there is one remarkable property of the -lattice which bears closely upon the present problem .
In general , two planes of the lattice which are nearly coincident in direction are by no means alike in other respects .
In the cubic lattice , for example , the cube face is the most dense , and the faces which approximate most closely to it in density are not the planes which most nearly coincide with it but the dodecahedron octahedron .
the ) or the by planes coincident with them is to replace planes of the greatest possible reticular ] lsity by planes of the least possible reticular density .
And yet the c1ystal is bounded by the latter rather than by the former .
Is it not possible that the supel.saturated liquid in immediate contact with the growing alum crystal consists of particles of alum uniformly mingled with particles of water , and that the act of crystallisation consists in the escape of the water , and consequent solidification of the alum ?
Immediately before the act of crystallisation the alum particles in the solution cannot be so closely packed as those in the crystal , since they are separated by particles of water ; if , then , they are to be laid down in plane layers , and are yet to find their places at once as constituent parts of the crystal structure , they will be laid down ] along planes reat reticular density but along vicinal planes .
Fig. 21 gives a crude representation* of a cubic structure in which the particles are so widely spaced immediately before the act of orystallisation that they solidif.y along the plane ( 025 ) .
The density in the cube face is times as great as the density in this plane ; for in the cube plane the mesh is a square ( side in the plane ( 025 ) the mesh is a rectangle whose sides are and respectively .
It is , of course , impossible from the constitution of a single layer of particles to deduce that of the solution in the neighbourhood of the crystal as a whole ; and the refractive index presumably only gives the average constitution of a certain Iayer of the solution , thin though that layer may be .
Nevertheless , the figure indicates that , if the crystallising particles be regarded as a shower falling normally upon the surface , the shower is so dense upon a growing cube face that it would leave little spaee for the escape of the solvent , whereas upon the face ( 025 ) , although equally dense as measured along lines perpendicular to the plane of the diagram , it .
is more than five particles are only 1epresented as cirdes in contact with each other , in order to make the structure more clear .
{ , OF IN CRySTALS .
times iu ) of the With vicil 1 the sort actually , the is , of course , very greater ; ) vicinal plane ( Okl ) in the cubic lattice inclined at to ) cube face has rectangulal Inesh whose sides are and 1 ( about ) , and the density in the cube face is therefore about 114 times that in the vicinal plane .
] Although the plane ( 025 ) has a open structure , the successive layers of parallel to this plane are far more closely packed than successive cube layers , as is clear from the figure ( lines 1 , 2 , 8 , 4 ) .
The black circles represent the particles which , at a given moment , bound the crystal ; the plane which preceded it is } resented by shaded circles whose centres lie on the , and the set of particles in the act of is represented by dotted circles .
If we are to speculate concerning the arrangement of the material in layers further removed from the surface of the crystal , we may perhaps suppose that along lines normal to the surface ( the arrows of fig. 21 ) the particles are even as closely packed as along an edge of the cube .
Each successive layer is then formed as the crystal solidifies by the particles immediately behind those of the newly solidified layer slipping sideways into their places .
Further , in the case contemplated , each dotted circle in the figure may be taken to represent a line of particles perpendicular to the plane of the drawing , which are packed as closely as the particles along a cube edge in the crystal , i.e. , as closely as the horizontal lines of ) articles .
The act of crystallisation will then consist in the deposition of such lines all parallel to the cube edge , but so widely spaced as to lie in a vicinal plane ( Okl ) .
This corresponds with the fact that the vicinal planes , though vary in inclination , always belong to well deiined zones ; in this instance they will lie in the cube zone , and sodium chlorate possibly affords an example of vicinal planes of the sort depicted in fig. 21 .
The determinations made above enable us to compare the weight of material contained in a given volume of the crystal with the weight of the same material in the same volume of the solution ; that is to say , density of the substance in the crystal with its density in the immediate neighbourhood of the crystal , ignoring the solvent .
Thus for Taking the specific gravity of the crystallised salt to be , then 100 cub. centims. of the solid will contain 17 grammes of ; and 100 cub. centims. of a solution containing per cent. alum , and of specific gl.avity , will contain grammes of alum .
The density of alum in the crystal is , therefore , about times that in the adjacent solution .
For sodium chlorate :Specific gravity of the crystallised salt cub. centims. of the solid will contain ralnmes of sodium chlorate ; and 100 cub. centims. of a solution containing per .
of the salt will contain rammes of sodium chlorate .
The density of sodium chlorate in the crystal is , therefore , about times that in the adjacent solution .
For sodium Specific gravity of the crystallised salt cub. V0L .
CCII .
PROFESSOR A. MIFRS : AN ENQUIRY INTO of solid Jlles of NaNO ; 100 cub. centims. of a cent. solution contain ] Ilnnes N The density of sodium in is , efore , about times that in the adjacent solution .
thing more concerning the nature of the liquid , it is difficult to relation between these numbers and the particular vicinal face If assume that not only the particles in the layer of liquid in immediate contact with the face , but also thosc in consecutive layers for an appreciable distance , have the arrangement as in the face itself ; then , if the consecutive layers in the lrated by a , and if be the area of the elementary in the grow i face , constant volume of the elementary parallelepiped in the lattice .
And if consecutive layel .
S in the liquid be separated by an interval , then the volume of the elementary parallelepiped in the liquid is The ratio of the density of the rial in the crystal to that in the liquid is therefore , and in substances considered above this quantity would be from to 3 If the relative distances be calculated on the assumption that be , as suggested above , not less than the least distance which separates adjacent particles in the c1ystal , the numbel .
S obtained for vicinal faces inclined at 30 ' to the octahedron are not with these , but are more than five times as great , whether the structure be the cube , the centred cube , the cube with ce1ltred faces ; and for vicinal faces inclined at only to the octahedron the disproportion is far greater .
It will be iced , however , that to assume any uniform distribution of the material in the liquid is to endow it with a crystalline structure before it solidifies .
There is , perhaps , ) othing impossible in the idea that the material may be already arrallging itself immediately before the act of crystallisation .
( There is , , no evidence that this is the case .
) these conditions , the solution could not possess the properties of a liquid , and , inter , might be exlJected to exhibit double ; it was with this possibility in view that the rongly bgent sodium nitrate was chosen above as a substance to be upon , in the hope that evidence of birefringence might be detected in the olution in contact with the growing face ; but no such evidence could be found by the nlethod of total reflection within crystal .
obtain like a correct estimate of the elative sizes of the meshes in the vicinal faces , and therefore of the distance which separates consecutive layers parallel to them , it would be necessary to measure their angles accurately not to minutes , but to seconds .
To determine from the constitution or specific gravity of the liquid any relation between the arrangement of the particles in the vicinal face in the solution , some assumption must be made concerning the successive layers in the latter .
VARIATION OF ANGLES ) IN TALS .
Such speculations are , however , at present beyond the test of is it possible to say what particular faces correspond to a certain of the solution until more is known concerning the nature of these layers of solntion immediately adjacent to the crystal .
In particular , it will be necessary to make accurate determinations of the change of volume ) luced by dissolving the crystals .
It is sufficient to point out that the widel spacing of the material before it solidifies is a reason why the crystallising substance should descend the crystal in a shower which is not very .
and this is a the growth should take place al viciJlal planes , and not along with simple indices .
That the growth does take place along anes has been abundantly proved , at any rate for lum , in the preceding pages ; that the spacing is very wider in the liquid than in the crystal has been proved showing that the liquid about to crystallise is scarcely more than saturated .
Very slight variations in the ration of the solution would probably correspond to considerable differences the vicinal planes produced .
The obseryations ) raise mally ] questions ; I have thought it better to confine lllyself in ) lmainly to ) , and to reserve further tion f another occasion , ) the the enquiry into the nature of the liquid ystal , ] opened , ill be pursued by in a xhaustive 1 Note.\mdash ; A brief statement of some of the esults recorded in I. .
) ) was published in the Repolt of the ]itish A for 18 , p. In the latter of the research I have received useful ) ftnt , .
RICHARD , by whom of the observations Il .
lverlnade .

