THE FAMILIAR right angle prism turns I rays through an angle of 90°, inverts the image, but does not reverse it. The elbow, telescopes of military gunsights on the other hand, require a prism that will similarly turn rays through an angle of 90°, similarly invert, but will also reverse the image. In actual application of course, such a prism is called on to normalize the already inverted and reversed image produced by the gunsight's telescope objective.

The common right angle prism can be modified so that it will do all these things. Russell W. Porter briefly explains this modification thus:

"Let us start with the familiar right angle prism A, Figure I, top, indicated with light lines. We grind away two sides of its hypothenuse face until only the line BC is left. This gives us two faces, D and E, which we shall call the roof. The sides of the roof must make an angle to each other of 90°.

"Now consider a beam of light entering the prism from F and emerging at G, at right angles to the immergent beam. If two arrows are used to represent the incoming object, it will be seen that, on emerging, the image has been reversed, as the arrows show.

"By tracing the arrow points and tails through the glass, shown by dotted lines, their rays are seen to cross each other by internal reflection from the roof faces. The image is also turned through an angle of 90°."

It may prove difficult to visualize this from a drawing, if the reader is not already familiar with roof prisms-it usually is confusing even to those who, for the first time, actually examine a roof prism. At first one looks very complicated. An added reason for this confusion is the fact that in such a prism there apparently are nine facets. Five of these are not, however, true optical facets The two tips of the prism which lie outside the circle of rays that reach the entrant face, as well as the heel part of the right-angle portion which lies outside the telescope's aperture, are ground off to save space in the mounting, and are left fine-ground, as are the two sides.

Thus a manufactured roof prism has only four polished, flat, optical faces. These are: the entrant face; the two roof faces which were made over from the right angle prism's hypothenuse side and which are at exact right angles to each other; and the emergent face at right angles to the entrant face.

G. B. Amici (1786-1864), an Italian, invented this prism, which is sometimes called the Amici prism. Sometimes, too, it is called other names not fit to print, by those who wrestle with its difficulties. Ordnance Document 1065, prepared under the direction of the Chief of Ordnance by Dr. I. C. Gardner, Chief of the Optical Inspection Section of the National Bureau of Standards, states: "It is one of the most difficult to manufacture because the roof angle cannot differ from 90° by more than a few seconds. Angles having the requisite accuracy cannot be . produced directly, but must be carefully approximated by usual manufacturing methods: after which the faces must be carefully polished by a skilled operator and individually tested until the angle is so nearly correct that the image is not doubled. An error of a few seconds is sufficient to make the prism worthless. This tedious method of production limits the output, as few men become sufficiently skilled to do this local retouching." Local retouching and "hand-correction," mentioned in the article for general readers elsewhere in this number, are the same thing.

Roof prisms are used in elbow telescopes, and elbow telescopes are used in all anti-aircraft sights, in order that the observer may look horizontally into the eyepiece. In the panoramic field-gun sight a horizontal element with a rotating objective prism of right angle type is added to the top of the elbow telescope, so that the cannoneer, without changing position may sight, first on his "aiming point" (not the actual target but some convenient fixed reference point), then turn off his azimuth, corrections for windage, drift, and so on, and fire the gun. This is the method of indirect laying of guns. That method is predominant because today the gun usually remains concealed and the firing data are given to its crew: by telephone from a distant observer who can see the target. This explains why a straight, direct, telescope, used as a sight, which would seem to be the answer that would do away very simply with the roof prism problem, could seldom be used.

Every amateur telescope maker will be curious to gain a rough idea of the high spots of the procedure of making roof prisms. To this end MacTavish has contributed the following, by invitation:

"In Figure 1, bottom, A is a side elevation of a roof prism, showing the 90° angle ('end angle') between the entrant (immergent) and emergent faces, also the 45° angle between the roof line and the entrant and emergent ('end') faces. In addition, the roof faces are 60° to the end faces.

"In producing the prisms, all angles except the roof angle are controlled in grinding and polishing to 2' of arc. This is assured if there is a light-tight reading under a fine standard square and a 60° gage. The roof angle is then brought to 90° +/- 2" by hand correction on a flat lap of pitch.

"The four optical facets must be flat within standard precision tolerances. Pits, scratches, grey surfaces, striae, ream, bubbles, nicks, and small fractures are not tolerated.

"Mechanical dimensions are held to close tolerances.

"The following roughly outlines the main processes, omitting the finer points, also the troubles-the former would add perhaps 25,000 words to the account, the latter from 50,000 up to, say, 1,000,000.

"The work starts with thick slab glass, though pressed forms are also used.

"The glass is sawed into rectangles, each large enough to make two prisms when later sawed diagonally.

"Groups of these are cemented to flatiron plates and held against wet abrasive grains on round, rotating, flat, iron plates, (Figure 2) first on one side, then the other, till they mike correct to the prism thickness dimension specified.

"Several of these squares are cemented together in stacked-up form, and one side 13 and one end of this block are ground square to the sides and to each other.

"The same cemented group is cemented to an optically flat iron plate with hot wax and the other side and end are ground parallel to the first side and end, respectively, in several operations.

"The pieces are removed, uncemented, and each square is cut diagonally with a r. diamond saw.

"The sharp tips are then ground off and left fine ground permanently.

"The 90° angles are corrected by hand with fine emery; otherwise the roof-face-to-end-face angle will depart grossly from 60°, and the roof-line-to-end-face angle will depart grossly from 45°. Other sources of error at this point must be skipped.

"Next, the two roof faces are ground on. The prisms are inserted in accurate metal V-blocks (Figure 3), with V's to hold them laterally, also notched metal inserts to hold them longitudinally (these are shown disassembled in Figure 2), and the roof faces are ground on, first one face of all, then the other face of all.

The roof faces are miked uniform all around through the V-block, so that the roof ridges will be in the exact center of the prisms.

"After the roof faces are ground on, the 60° angles are corrected by hand retouching against an iron plate with fine abrasive-a very ticklish, fatiguing task.

"Next, the sharp edges of the prism are beveled.

"The prisms are again cemented, bottoms up, into another V-block without inserts and the small (non-optical) bottom faces are ground off.

"Preparation for polishing is begun. Face by face the prisms, somewhat separated, are cemented to an optically flat iron plate with melted paraffines or cold oil. A ring of sheet metal is placed around this plate and special plaster is poured around the prisms. This forms a matrix, or round block. The ring is removed, the bottom iron plate is warmed (if paraffine is used), and the block slid off. The plaster is cut back 1/16", so that the prism faces stand slightly in relief, and is waterproofed with shellac.

"This prepared matrix (Figure 4)- two of these also are visible in an illustration on page 203-is now like a big

round blank with which you would make a flat. It is so regarded and dealt with in polishing, figuring, and testing. Another flat, good to a tenth of a fringe, is used for testing it. Here there is, however, a complication-for plaster expands on setting, also changing with temperature changes, so that it is not a perfect matrix. This often causes much trouble. One must also learn to seat the prism so that errors will not be introduced to spoil the already corrected angles.

"Polishing is done face up on a modified Draper machine. The polish must be free 3 of all defects-clear and grainless.

"The two end faces are polished first, successively, then one roof face.

"Before the second roof face is blocked, the roof angle must be corrected by hand grinding as closely as possible to 90°, so that when the prisms come out of the block, the angle error will be 2' or less, in order to save time later in hand correction.

"Great care and correct abrasives must be used in grinding the final roof side. No tiny chips, nicks, or fractures are permitted along the roof line. Even No. 1000 Carbo will cause fractures here, and superfine emeries are used.

"The second roof face now is polished (Figure 5).

"Thus all four optical faces have now been polished, and next comes one of the most thrilling operations known to me in optics-correction of the roof angle by freehand polishing on a small flat lap of pitch. Correction of the roof angle is not in itself so difficult, but the lap must be held flat at the same time-flat within precision tolerances - without astigmatic fringes or turns. The crux of the operation is the preparation of the lap and the use of strokes which will produce the desired results-and, of course, a skilful man doing it. Such a man, one of the most skilled hand-correctors in the country, is Pavel Uvaroff, who has corrected as many as 55 roof prisms in a single day. Here is work where money and fancy equipment cannot substitute for the man. Figure 6 is Porter's sketch of one of numerous differing finger positions favored by prism correctors.

"As the angle is being corrected, frequent testing is necessary. A target of crossed thin lines is placed about 30' distant and its image is viewed through the prism by means of a 20X telescope. The vertical lines do not double unless astigmatism is present in the prism. The horizontal lines double to a greater or less degree, depending on the angle error. The aim is to correct the prism till they coincide and are as sharp and clear as the vertical lines. With greater magnification we have proved that an error of 1" of arc can be discerned.

"Readers who are already familiar with roof prism making will have found many omissions in the above hop-

skip-and-jump outline. This is due to space limitations. I was invited to touch only the high spots, with the aim of giving the average amateur telescope-making reader a generalized picture.

"One noticeable difference between mirror making and roof prism making- noticeable because, if the maker doesn't notice it, the purchaser's inspector certainly will-is the fact that the maker is no longer his own judge and jury. Any tendency to be optimistic about one's own work is cancelled out by the same cruel and heartless inspector. The prisms must equal specifications or else."

BECAUSE it is anticipated that some readers will wish to go in for roof prism making, it probably will be a kindness-nothing less-to state here with entire candidness that it almost certainly is. now too late to begin. Tooling up, alone, requires weeks or months, while learning the difficult art usually adds from three to six months more. It is not now believed: that this would result in much, if anything, more than a bad headache, irretrievable expense, and disappointment. In any case, the Scientific American Roof Prism Making Program is no longer open to new entrants-it has done its work.

If this seems to be bad news, you may be tempted to ask why you were not invited to participate at the beginning, 18 months ago. Every effort was made to locate the ablest men, short of undesirable publicity, simply because every such man was actually wanted. Yet, despite a variety of under-cover fishing expeditions, many good men were no doubt missed.

Requirements for admission were: several mirrors previously made, also flats. This led to some heartaches-the less the applicant had done, apparently the greater the heartache; in fact, several who applied without having made any mirrors at all nearly died of heartache. This seeming (and actual) exclusiveness arose from only one motive: the leaders found it impossible to service more than approximately 80 candidates. As a war was on, it was found necessary to be arbitrary-for which apology and the above explanation are now offered.

Even as it proved, however, the leaders who tried hard to coordinate the work will probably never be the same again. Your scribe, after 18 months of preoccupation with the job, has selected a secret hole, acquired an eight-volume "History of the Dark Ages" (in which wars were fought without roof prisms), crawled in, and is about to pull the hole in after him -Goodbye!

Emerging after a spell in the Dark Ages, he will, however, if the war situation or its end permits, endeavor to offer a second account in which details of the: program's accomplishments can be discreetly revealed. This will include a statement of the true identities of the heroes of the preceding story, each at present protected from sabotage under a name that, by censor's request, is fictitious.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.