]\gt ; Mr. .
H. S. Jones .
The Flexure of Telescope [ Apr. 26 , 7 .
Conclusions .
( 1 ) Recombination appears to take place according to the simple law or more generally ( 2 ) The coefficient of recombination with the pressure and throughout a considerable range of pressure is proportional to the pressure .
In conclusion I desire to express my appreciation of the kindly interest which Prof. Sir J. J. Thomson has shown throughout this work .
The Flexure of Telescope Mirror-discs Arising from their Weight , and its upon Resolving Power .
By H. SPENcElt JONES , B.A. , , Jesus College , Cambridge , Isaac Newton Student .
( Communicated by Sir Jossph Larmor , F.R.S. Received Apri126 , \mdash ; Read June 19 , 1913 .
) The resolving power of a reflecting telescope is proportional to its aperture , the mirror being supposed accurately a paraboloid of revolution , sa a bundle of parallel rays pass through a geometrical focus after reflection .
Regarded from this point of view alone it is of advantage , in the construction of a reflecting telescope , to make the mirror of as large a size as the mechanical difficulties incident to the construction of large mirrors will allow , difficulties , that is to say , such as that of obtaining the glass of the necessary homogeneity throughout , in order to avoid distortion owing changes of temperature , and the difficulties of grinding , shaping , and polishing .
There is , however , another factor to be considered which affects the problem .
The mirror , even if perfect in other respects , will , when partially supported , be distorted by its weight , to a greater or less extent .
according to the nature of the support .
This distortion will cause a diminution in the resolving power , the ifect of which will evidently increase with the size of the mirror , and so will tend to counteract the advantage from the larger aperture .
It is conceivable even that there will , for any 1913 .
] Mirror-discs from their Weight .
given mebhod of support , be a critical aperture , an increase of the size of the mirror beyond which will actually produce a decrease instead of an increase in the resolving power .
The object of the presenb paper is , firstly , to calculate the nature and amount of the distortion which is produced by the weight for various methods of support , and , secondly , to investigate to what extent this distortion will affect the resolving power of ths instrument , and whether any limitation is thereby placed upon the size of the mirrors which are likely to be practically attainable .
The types of support which are considered are necessarily comparatively simple and somewhat ideal : the difficulties of the mathematical analysis impose these limitations , but it will be seen that they are sufficient to enable us to give a definite answer to the problem as to whether the critical size of aperture is one which is likely to be reached in the construction of large mirrors , or as to whether this critical size is so large as to be of no practical significance .
Since the equations of elastic are ]inear it is sufficient to consider the flexure of the disc for the two cases in which it is horizontal and vertical .
The case in which it is inclined at any angle can then be obtained by a combination of these .
The notation used hout will be that used by Prof. Love in his ' Mathematical Theory of Elasticity.'* We will discuss first the case in which the plane of the disc is horizontal , and will treat it as a circular plate of uniform thickness .
This is nearly true for large mirrors and will be a sufficient approximation for the present discussion .
We consider several different supports .
( 1 ) The disc is held in a horizontal position , being supported at its centre , its edge being free .
We obtain the solution by a combination of various stress systems , which together are to the correct body forces and are to be so adjusted as to satisfy the terminal conditions .
The conditions in this case are that the normal displacement at the centre should vanish , and that the stress resultants , and stress couples , should vanish at the edge .
We use the following systems of stress\mdash ; A system in which all the stress components vanish except , being the weight of the plate per unit volume , and the axis being vertically upwards , and the origin in the central plane .
2 is the thickness of the plate .
*Camb .
Univ. Press , 1906 , 2nd Edition .
Love , loc. , S 294 .
Mr. H. S. Jones .
The I ?
exure of Telescope [ Apr. 26 , The corresponding displacements may be taken to be given by Also all stress resultants are zero .
The axes of and are in the plane of the plate ; are the components of displacement parallel to the three axes ; , as usual , denote Poisson 's Ratio and Young 's Modulus of Elasticity for the material of the plate .
A stress-system in which there is a pressure 2 on the face of the plate .
The solution for a plate bent by a uniform pressure over one face is given in Love , S307 ; we must in the solution there given put and so obtain whxy .
These give , in the central plane , a radial displacement of amount , and a nolmal displacement where is the igidity of the plate .
If the radius of the plate be , we have also at the edge 1913 .
] Mirror-discs Arising from their Weight .
( ) We need other solutions which give no resultant body forces in order to adjust the boundary conditions .
We may take giving For this system the normal displacement of the central plane vanishes , but there is a radial displacement in the plane of amount We finally take the solution corresponding to a system of generalised plane stress , given in Love , SS303 and 304 The components of displacement in the central plane vanish .
If is the normal displacement of the central plane it is shown that satisfies where is Laplace 's Operator in two dimensions , and that the stress resultants are derived , when is known , from the formulae . .
We have given by an expression of the fornl since vanishes at the centre .
Therefore Combining these four solutions , the conditions that vanish at the edge give .
The total value of at any point is i.e. there is a downward displacement at any point of amount Mr. H. S. Jones .
The Flexure of Telescope [ Apr. 26 , Putting 2 , the weight of the plate , the downward deflection at the is given by .
The second term in this expression is small compared with the first , the relative magnitudes being of the order , which is comparatively small , since the usual order of magnitude of the thickness is 1/ 10 of the aperture .
The correctness of the first term of this expression oan be proved by the use of the approximate theory which holds for the flexure of plates .
This approximate theory is very useful in obtaining solutions for more complicated cases of support , when the more exact theory would either be very tedious or else impracticable .
It is shown by Love* that for a plate slightly bent by transverse forces only the stress couples at the edge are given by , where denotes differentiation along the outward normal at the edge , and the edge , is the radius of curvature .
If is the acting force per unit area , is determined by We therefore have for the case of a circular plate under the action of its own weight , ; that using the facb , where necessary , in the integration that vanishes at the centre of the plate : so that when we must have , since vanish at the boundary , and ) These values of agree in magnitude with those previously obtained , provided that the term in which involves be ected .
We therefole obtai the deflection at the edge iven by the approximate expression .
, S 313 .
1913 .
] Mirror-discs Ari.sing from their Weight .
( 2 ) We now take the case in which the plate is held horizontally and clamped at the rim .
Then we must havs both and vanishing at points on the rim ; and the radial displacement in the plane of the disc must also vanish on the rim .
If we combine the system of solutions above , the resudtant radial displacement is and so vanishes at all points .
We may take for the solution of in in this case Since is not infinite at the centre , we must have , and in order that may not become infinite at centre , we must have Thus the conditions at lire that and the depression at any point is given by , the central depression being In this case the exact theory agrees with the approximate theory.* The reason is that no terms upon are involved .
( 3 ) If the plate is supported at the edge , we must have vanishing there .
Ths first two evidently do so .
We can take the same solution for as in ( 2 ) , and the remaining conditions give and , giving for the value of at any point , a downward displacement , * See solution , in Love , p. 467 ( ii ) .
Mr. H. S. Jones .
The Flexure of Telescope [ Apr. 26 , The first term agrees with that obtained by means of the approximate theory .
Combining these three results we have , for the relative depressions of centre and rim , ( 1 ) When supported at the centre , .
( 2 ) When clamped at the rim , .
( 3 ) When supported at the rim , .
Neglecting the small second term , the ratios of these expressions depend only upon the value of .
If we put , which is approximately the case for glass , we have the relative deflections in these three cases given by the ratios 93 : 15 : 63 .
The deflection is thus much smaller in the case of the clamped rim .
( 4 ) It is to be expected that the more complete the support at the back of the mirlor the less will be the effect due to the flexure .
We will accordingly consider the case in which the plate is not only supported at the centre , so that there , but is , in addition , supported at the edge .
A sufficiently close value of the flexure is obtained by using the approximate theory , and since the work is much shortened by so doing , that is the course here adopted .
We have Thus Since at the centre , we must put ; and since at the edge , we obtai1l Also must vanish at the edge .
The second boundary condition is accordingly 1913 .
] Mirror-discs Arising from their Weight .
From these two equations we obtain and the value of at any point of the plate is given by which vanishes\mdash ; as it should\mdash ; when ( 5 ) We consider finally , as an example of a still more complete support , the case in which the disc of radius is supported at the centre and the edge , and also round the circumference of a concentric circle of radius We must now assume different solutions which hold according as If we may assume If we take We have here six c , onstants Ting .
Ihese are uniquely determined from the six boundary conditions , which are that , should vanish when when when , , , The condition that when has been satisfied already by our choice of From these equations we deduce that , Mr. H. S. Jones .
The Flexure of Telescope [ Apr. 26 , and also that These equations suffice to determine the deflection at any point of the disc in this case .
We have still to consider the solutions of the stress equations which are appropriate to the case in which the disc is held vertically .
The body forces are now in the plane of the disc , and there are no normal components .
We take the and axes in the plane , the origin being at the centre , and the positive direction of the axis being vertically downwards .
We will consider several simple solutions , by a combination of which the solutions appropriate to various boundary conditions may be obtained .
We take the simple solutions of the state of plane stress:\mdash ; If we assume we have a system stresses which give a body force parallel to the axis of amount , the weight per unit area .
The remaining solutions will , therefore , be such as to give no resultant body forces .
From these values we deduce the following particular solutions for the displacements , and the stress resultants are -why , which give and , etc. , all zero .
( ) A particular solution which gives no body forces is one in which the stress resultants and the components of displacement are iven by and 1913 .
] Mirror-discs Arising from their Weight .
and the rest zero .
Thus ( ) There are also a series of particular solutions for forces in the plane of the disc which are given by ' where is any plane harmonic nction .
These values give If we take ) , a plane harmonic of the third yree , we obtain \mdash ; cos2 and then the displacements are If the plate is held at the centre there must be a force applied there which is equal in magnitude to the weight of the disc .
The simplest case of a singularity arising from the application of a force at a point inside the body is given in Love , S148 .
The components of displacement are A and the stress components are given by , ? .
These are easily shown to give Mr. H. S. Jones .
The Flexure of Telescope [ Apr. 26 , ( 1 ) We combine these solutions by adding to the first , the second multiplied by a constant times the third , and times the fourth .
Then the stress components are found to vanish all round the edge provided that l , The coefficient in the stress resultants in the solution is then .
The resultant of the tractions on the surface of a small cavity taken round the point of application of the applied force must be equal and opposite to that force , and therefore independent of the shape of the cavity in the limit when its dimensions are taken very small .
The resultant is easily found to be parallel to the positive direction of the axis , and is therefore equal in magnitude and opposite in direction to the weight of the body , and so is correct .
The solution thus obtained is therefore the correct solution for the case in which the disc is supported at its centre with its edge free .
The displacement components are given by const .
The constant occurring in the expression for has to be determined from the condition that should vanish when vanishes ; but on account of the logarithmic term becomes infinite at the centre .
The reason for this is that the force cannot be applied at an exact mathematicai point : it must be distributed over a small area , and the value of the constant will depend upon the mode of this distribution .
The difference of the displacements in the direction of the force at any two points on the disc can , however , be uniquely determined by this formula .
( 2 ) If the disc is clamped at the and held vertically the displacements there must be zero .
To obtain the appropriate solution we combine the cases , ( ) , ( ) above , determining the constants introduced by means of the boundary conditions ; we obtain const .
1913 .
] Mirror-discs Arising from their Weight .
We choose so that the coefficients of in the expression for are equal , and so that vanishes , i.e. ' .
The constant in is so chosen that is zero on the rim .
We thus obtain ' and the corresponding stress components are easily shown to be which give the correct body force , and so all the conditions are satisfied .
The displacement at the centre is given by This displacement is entirely in the plane of the plate and is small compared with the displacement in the direction 11ormal to the plate , when it is horizontal , for in the latter case there is a term involving in the denominator .
( 3 ) For the case in which the disc is held vertically in a groove at the rim in such a manner that the normal displacements there are zero , but tangential displacements are possible , the boundary conditions are that the normal displacement and the tangential stress resultant must vanish at the rim .
Our previous solutions cannot be combined in any manner so as to satisfy these conditionsi .
We can obtain another solution for the case in which there are no body forces in the following We may express the stress components in terms of a mcCion , which satisfies , in the form .
We try a solution Mr. H. S. Jones .
The Flexure of Telescope [ Apr. 26 , and then we have cos3 , whence we deduce that , , , whence These three equations are satisfied by The normal component of the displacement at any point is We combine this solution with the solution above , for which the normal displacement at any point is Then , when , by putting , and equating the coefficients of to , we obtain ' 1913 .
] Mirror-discs Arising from their Weight .
so that the components of displacement at any point of the disc are iven by .
The central deflection is given by ' In the case in which the plate is inclined to the vertical there will be a superposition of the two effects .
There is now a component of gravity the plane of the plate and a component normal to it .
This produces the normal which is , in general , the chief distortion .
The component of the force in the plane of the plate produces small displacements in the plane , .
accompanied by very slight normal displacements , which are symmetrical about and vanish on the central plane .
We can obtain some information as to the tive magnitudes of these effects from the results which have been already obtained .
When the plate is supported horizontally at the centl'e it has been that the deflectio11 at the is .
Ths ratio of the two terms which occur in this expression is If we put , and , which gives ratio of aperture : ness equal to 10 , we obtain ratio of the two terms to be 1 : , so that the second term is tolerably small when compaled with the first .
When the plate is supported ab the rim the deflection at the centre is , and the ratio of th second term in this expression to the first mlder the conditions is 1 : VOL. LXXXVIII .
Mr. H. .
Jones .
The Flexure of Telescope [ Apr. 26 , the case in which the plate was held vertically and clamped at the edge we have seen that the central displacement is and the ratio of this to the small term occurring in the normal displacement when supported at the centre is and so is very small indeed .
The displacement in the plane of the disc when this is held vertical is in this case only about 1/ 16,000 of the normal displacement at the edge when the disc is supported horizontally at the centre .
If the disc is vertical and held at the centre the relative displacement for points on the same radius at distances from the centre is If cm .
and cm .
, then has a maximum value when , and this maximum is approximately which is rather larger than the small term ) which occurs in in the first horizontal case considered .
When the disc was vertical and supported round the rim the central deflection was .
This is about 5/ 4 the value when the edge is clamped .
It follows that , unless the disc when horizontal is very well supported at the back , the displacement normal to the disc which is produced by its weight is very large compared with the displacement in the plane of the disc when it is held vertically .
We have now to apply the results which have been obtained to discuss the extent to which resolving power is affected by the flexure .
We suppose the undistorted mirror to be accurately paraboloidal with a focal length and take a system of rectangular axes such that the axis is the axis of the mirror , and the vertex is at the origin .
Then the equation of the paraboloid is 1913 .
] Mirror-discs Arising from their Weight .
A bundle of rays all parallel to the axis will after reflection all pass through the point of intersection of the axis with the focal plane , i.e. through the point We consider now another parallel bundle , each ray of which makes a small angle with the axis .
The ray which passes through the point of the surface has the equation Then the reflected ray may be represented by Since the direction cosines of the normal to the surface are proportional to , and since the incident and reflected rays make equal angles with the normal , and are all in the same plane , we have the conditions ; and also together with Since has been assumed so small that its square and powers may be neglected , and since also we obtain for the values of the expressions , correct up to terms including and , to the same order of approximation , the co-ordinates of the point in which the reflected ray cuts the focal plane are given by or the distance from the axis of the point in which the reflected ray through cuts the focal plane is For a iven value of this is greabest when , i.e. for the point in the Mr. H. S. Jones .
The flexure of Telescope [ Apr. 26 , plane through the axis which contains the direction of the bundle .
For points in this plane the greatest value is at the edge , and is where is the edge value .
Since even for points on the rim will be comparatively small we may take this distance to be approximately Now two points of light , one of whioh is on the axis and the other at a small angular distance be resolved provided that where is the wave-length of the light , and is the semi-aperture of the telescope , for when this condition is satisfied the angular separation of the spots is of cient magnitude to prevent their first diffraction maxima overlapping Hence if two stars are to appear distinct , the breadth of the central image for parallel rays must not exceed Suppose now that the undistorted surface has for its equation Then , the components of the displacement of the point , are functions of and , and the co-ordinates of the displaced point are given by Since are all small we can in them put , so that they become functions of and .
The equation of the deformed surface is therefore .
We consider first , for simplicity , the deformation which is produced when the disc is horizontal .
In this case , the only displacement being norrnal to the disc and symmetrical , so that we can proceed in two dimensions .
The section of the surface by the plane has for its equation .
The equation of an incident ray parallel to the axis is , and since the ( lirection cosines of the normal to the surface at the point are proportional to it follows that , the reflected ray being and are given by the equations 1913 .
] Mirror-discs Arising from their Weight .
and the reflected ray is If , i.e. if the mirror is undistorted , there is perfect focussing at We suppose then that the nearest approach to focussing when the mirror is bent is on the plane , where is a small quantity which has to be determined .
Then the corresponding value of is given by where has been used to denote ( w/ \amp ; ) , and where squares of have been neglected .
On expansion , retaining only first power of , there results ( 1 ) In the case when the diso is clamped at the edge it has been shown that , where By differentiation of the above value of it is found that the condition for a stationary intercept is where , and then we get .
The condition for a stationary value becomes in the case at present under discussion , For a given value of there are two values of .
In the case in which , these combine to give an inflexional tangent at The value of at any point is The part which is independent of vanishes at the two edges , and the two parts add as shown in the diagram .
Mr. H. S. Jones .
The Flexure of Telescope [ Apr. 26 , The position of the stationary points depends upon the value assigned to and the values at the ends may be either greater or less than the maxima .
The most favourable case , the least breadth , will occur when the two are equal .
Retaining only the most important terms the maxima are given by and to the same order are equal to The end values when are equal to Hence the condition that this is equal to the maximum is which is satisfied by , giving a breadth of image 2 , equal to 4 , and this gives the limiting condition that The following approximate mean values for glass are sufficient to determine the magnitude of the permissible limits for , viz. :\mdash ; C.G.S. units , cm .
, and Putting , the condition is that , or If the mirror is so constructed that , then the limiting value of is cm .
, and so in this case the resolving power is lost for a very small breadth of aperture .
Since the limiting condition is const .
this result is also an example of the general law that the thicker the mirror the larger is the permissible breadth of the aperture .
1913 .
] Mirror-discs Arising from their Weight .
In the case when the disc is horizontal and supported at the , the value of was found to be ( say ) , where has the same value as above .
Proceeding exactly as above it is easily found that the stationary values are given by : and when The maxima and minima are again , and so the condition that these are equal to the end values is which is satisfied by a breadth equal to exactly as before , and so\mdash ; since has the same valne\mdash ; the limiting value of in order that the resolving may still increase with increase of aperture has the same as before .
( 2 ) Coming now to the nex6 case considered , in which the plate was supported both at the ce1ltre and at the edge , the value of was shown to be where were determlned , and here Then the intercept on the plane is approximately , ectin g all terms of higher order .
The maximum value of the intercept is determined by the condition and for this value of If we adjust so that the maximum and the value at the are equal and Mr. H. S. Jones .
Flexure of Telescope [ Apr. 26 , opposite , then the from the two halves of the mirror overlap .
This requires that where the value of determined by the maximum condition has been used .
Since the focal length is usually about five times the aperture , i.e. the ratio of the terms is , and so very little error will be caused by working with the approximate first terms .
Then for the determination of the value of at the maximum we have the equation Now , approx. , so that ; or , putting By trial it is found that nearly satisfies , so that we write where is small .
A second approximation gives , so that it is sufficiently acourate to take at the maximum .
Since the first terms are now comparatively small , we include also the next order terms on substitution of this value of , and by writing , and it is found that the breadth of the image is 2 therefore the limiting condition is that or , using the same values for the constants as before , this condition ives , and if cm .
, or inches , and so the limiting permissible value of the radius , in this case , the thickness being of the aperture , is about 4 feet .
1918 .
] Mirror-discs Arising]Weight .
If we use a thicker plate we can use a mirror of aperture , e.g. if , then inches , or the permissible radius is nearly 7 feet .
By comparison with the preceding cases , it is at once apparent how great is the advantage which is obtained simply by supporting the mirror at the centre in addition to the edge .
( 3 ) It now remains to discuss the last mode of support considered , when the disc was horizontal , viz. , when it is supported at the centre , round the circumference , and round a concentric circle of radius the values of the displacement which were found for the two portions of the disc we obtain for the values of the intercept on the plane the values , according as or respectively , Jog which are continuous at , as they should be .
We can abbreviate the work and at the same time illustrate the advantage given by the additional support if we take a special value for .
The simplest value to take is , and then , by putting , we have givetl by and also On investigation , it is found that the conditions cannot be satisfied by supposing the maximum deflecLion to fall in the inner portion , and by making it equal to the end dellection .
We therefore make the maximum in the outer portion equal to the end deflection .
The condition for the maximum is and the corresponding maximum is .
H. S. Jones .
The Flexure of Telescope [ Apr. 26 , and the equation to determine is ab On substitution of the numerical values , we get 2.02 By trial it is found that nearly satisfies , and by substituting , and retaining first powers only of , a nearer approximation is .
Including now the next order the total breadth of the image is found to be 4 ak , and the limiting condition is that , so that , or cm .
inches , if The limiting value of the radius of the mirror is thus increased to about 10 feet , as compared with the 4-foot radius obtained in the case in which the support at half the radial distance is abssnt .
The cases which have now been discussed are sufficient to enable us to conclude that as far as the displacement normal to the plane of the disc , produced by the component of gravity in that direction , is concerned , the effect on the resolving power may be practically nullified if care is taken to support the mirror well at the back .
Although the methods of support considered above are necessarily somewhat ideal , it is certain that with a mirror well supported at the back , as e.g. is the case in the method devised by Ritchey for the large reflecting telescope at Mount Wilson Observatory , it is possible to use mirrors of an aperture which is far beyond the limits which at present are practically attainable .
This discussion has also shown how , for a given mode of support , larger apertures may be used provided that the thickness of the mirror is increased sufficiently .
The bhickness to be used must of course be determined by practical considerations , such as the possibility of obtaining a homogeneous mirror , or by the question of the weight which the supports are designed to bear .
For the remainder of the discussion we shall suppose the mirror to be well enough supported at the back for the normal component of gravity to have no effect upon the resolving power , so that it only remains to discuss the effect of the displacements whicf ] occur in the plane of the disc .
We will 1913 .
] -discs Arising from their Weight .
therefore consider the case in which the disc is vertical , as then these displacements will have their maximu value .
The nature of the back support does not now concern us , and whatever this may be the disc must also be held in some manner at the edge and it will be impossible to prevent the displacement in the plane of the disc from taking place .
Putting , the equation of the mirror becomes The equation of the normal at is where and it is easily shown that corresponding to a ray incident at the point of the surface parallel to the axis , the reflected ray is where are given by ; and as far as the first order only in the displacements , by substituting for in terms of , and putting , we have ; and , similarly , For simplicity of discussion , we will consider first the case in which .
we consider for the present only the rays which are incident at points on the mirror in the vertical plane through its axis .
For the case of the mirror which is clamped at the where and so , and Neglecting the term in and all terms of a higher order , we see that , for a given value of is a maximum when Mr. H. S. Jones .
The Flexure of Telescope [ Apr. 26 , We suppose that positive : then the intercept is positive and is equal to The value of steadily diminishes with increase of , since is supposed positive , and so the greatest negative intercept is from and is , so that the total breadth of the band is and to this order is least when we put A similar result is obtained if is supposed negative , since then the greatest negative from Thus the least value of the intercept on a plane perpendicular to the axis produced by the reflection of a beam incident parallel to the axis in the vertical plane through the axis is 3 .
The upper limit for in order that the resolving power may not decrease with increase of the size of the mirror is given by 3 or and if the focal is five times the aperture , , we obtain the limiting value of to be cm .
inches , and so the limiting radius of the mirror is about feet , and quite a lar .
size is permissible .
Whereas , in the cases previously considered in which the mirror was horizontal , the determining factor was the ratio of the aperture of the mirror to its thickness , the limiting value of the radius being found from an equation of the type constant ; in the present case the determining factor is the ratio of the aperture to the focal length , the limiting value of being found from an equation of the type constant .
If it was desired to construct a mirror of given focal length and of as large an aperture as possible , the limit for the aperture could be determined from the above .
Thus , if the required focal length were 50 feet , the largest radius which may be used is 280 cm .
, or about 9 feet .
In the more general case when is zero we have for a first approximation Or , since we have , Corresponding to the centre , we have , and the radius vector from this point to any other point of the patch is given by 1913 .
] Mirror-discs from their Weight .
and this attains its greatest value when and have their greatest values , i.e. at the two points on the circumference in the verlical plane through Che axis , and then .
This gives the 1naximum extent of the swing for light incident at any point , and is the case discussed above .
In the obher cass considered , in which the normal component of the displacement at any point of the edge vanishes , the normal component of the displacement at any point of the disc , using the values of 2 , previously obtained , is given by -S The poinl in which the reflected cuts the focal plane is iven by giving a normal component of amount , and , since , we obtain so that and so is given by 2 Or , , where is constant for Mr. H. S. Jones .
The Flexure of Telescope [ Apr. 26 , definite radial direction , the normal component is The variation of this expression with for a definite value of , i.e. along a definite radius , is and the radial component of the intersection with the focal plane is found to be , on substitution for There are two cases to be considered , according as If , the radial intercept continually decreases with increase of so that the extreme values arise from the centre and circumference , and are at the centre and at the edge , and the maximum extent of the displacement is the difference of these , and is very nearly .
This expression decreases with increase of and so has its maximum value when , i.e. in the plane , and is then .
If , then the radial intercept has a stationary value the value of given by and this value of will be less than provided that Thus if , the radial intercept continuously decreases from to .
If , however , it reaches a minimum for a certain value of which is less than and is given by the above equation .
After this it will commence to increase , and the extreme values are at the centre and at this minimum .
For this value of the intercept is 1913 .
] Mirror-discs Arising from their Weight .
and the maximum breadth of the image on the focal plane in the definite direction , , is Consideriug now the variation of this expression with , as increases from to , we find that it decreases , and so has its greatest value , viz. , This is less than the greatest when and so the limiting condition for is that , ' and on riting this gives a linliting value for of 358 cm .
, or 141 ches , so that the disc may have with this mode of support a radius of nearly 12 feet .
This method of support corresponds closely to that used by Ritchey , who , in addition to the back support which prevented the normal displacements from being appreciable , placed the mirror in a circular metal band supported on the outside .
In this way , can be no normal displacements at the edge , but there is perfect freedom for tangential displacements to occur .
It has been assumed in the discussion immediately preceding that mirror is held vertically .
The displacements in this case necessarily attain their greatest possible values , and so if the radius is less than the critical value for this case , it will be so any augle of inclination .
The resuIts obtained in the present discussion may be briefly summarised as follows:\mdash ; 1 .
The components of the displacement of any point of the mirror-disc arisin from its weight have been calculated for various modes of support , both in the oase when the mirror is he , ld horizontally , and also when it is held vertically .
2 .
It is found that in all cases the displacements in the former case are very much larger than the latter , and that if the mirror were not well supported at the back , the lesolving power would begiu to decrease as the aperture was increased for comparatively small breadths of aperture .
The more complete is the back support , the larger of course is the critical size of mirror , and with simple forms of support this critical size is far larger than is likely to be practically reached .
3 .
It is also shown thnt for any given mode of support the critical size of the mirror may be increased by increasing at the same time its thickness .
522 Drs. J. A. Harker and G. W. C. Kaye .
Electrical [ May 6 , 4 .
When the mirror is held vertically the displacements are small , but the support cannot be so arranged as to rid of them , as was the case with the normal component .
It is shown , however , that the critical size of mirror is comparatively large , very much larger in fact than any mirrors at present in use , the limiting radius being somewhere in the hbourhood of 12 feet in the two cases which are discussed in detail , when the telescope is so constructed that its focal length is five times its aperture and the aperture ten times the thickness of the mirror .
For a given mode of support a larger aperture may be used by increasing the focal length .
On the Electrical Emissivity and of Hot Metals .
By J. A. HARKER , D.Sc .
, F.R. , and G. W. C. KAYE , B.A. , D.Sc .
, National Physical Laboratory .
( Received May 6 , \mdash ; Read June 19 , 1913 .
) lntroductory .
In February , 1912 , the authols communicated to the Royal Society an account of some experiments on the emission of electricity from carbon at high temperatures .
* In this investigation all the experiments were conducted at atmospheric pressure , and some evidence was forward that under these conditions the carriers of electricity appeared to consist almost wholly of ' ' spnttered \ldquo ; matter , and that corpnscles\mdash ; the carriers of negative electricity in high vacha\mdash ; played in these experiments but a minor or at any rate an indirect part .
It was shown that , even in the absence of any applied potential , it was possible for charged particles to escape in sufficient number from a hot carbon surface to give rise , at temperatures , to currents of ammeter raf , her than electrometer magnitude .
The particles appeared to be emitted with considerable velocity and the emissivity of the radiating surface sending out the particles seemed to depend primarily on the temperature alone .
It was , however , influenced to some degree also by the nature of the surrounding gas .
Depending on this emissive property of carbon , a kind of ionic dynamo was constructed , capable of lighting intermittently a small group of glowlamps ( 2 volts , 3 amperes ) .
* Harker and Kaye , ' Roy .
Soc. Proc 1912 , , vol. 86 , p. 379 .

