There is another side of this method. We have just indicated refinement in detail of one kind. but the machine lends itself to analytical work not less than synthetical. In statistical investigation the analysis naturally becomes finer as the area enlarges, and here the sorting box is of great service. As has already been stated the cards are primarily massed in enu meration districts For such small areas, the information required groups the population under comparatively few heads. In practice it is found that such classification can generally be counted on the 40 dials that I he machine embraces normally as a full equipment ; and the arrangement is made accordingly. But while counting this classification, the cards can also be assorted into groups that will form the basis of the analysis for the next larger group of territorial areas ; so that if the cards are divided into twenty groups, we shall have at the next handling of the cards, a classification of 20 x 40. or 800 heads. If, at the next step, we subdivide each one of these twenty groups into twenty Ilion% the third hand ling of the cards will give us 20 x 20 x 40, or no fewer than 10,000 heads. Time a very few manipulations will give an extraordinarily fine degree of analysis, and the compilation will have a value from its minuteness that could be reached in no other way.
Added to the ability to secure special details, finer analysis, and the economy in time and labor, we have the greater accuracy, '1'lle machine automatically throws out any card that is wrong. Suppose, for inst :owe, f hat age or sex has not been punched. Where there should be a hole for the plunger-pin to go through, closing the circuit, the card is intact. The cireuit is open, and the inonitor bell just to the left of the press, refuses to give its cheery signal of correctness. It is then a very easy matter to refer Inick to the schedule stowed away in the old church across the street, and fill up the deficiency by the paradoxical pro cess of making a hole. Supp15se it was desired to connect up the machine so that only cards for New York should be counted. A mis sorted card belonging to Chicago would at once be rejected. The gang punches of the two cities not agreeing, the wrong cards would leave the circuit °lien.
That all of a batch of cards purporting to represent some one class are properly assort ed, is simply ascertainable by passing a wire or needle through the holes representing the given class: This could evidently not be done with written canls, and locating a mis placed written card among a million other cards is practically impossible. The proba
bilities of error in reality narrow themselves down to the punching, and even then the only errors that escape detection are those iu which the information given, while it may not furnish the exact fact, is still consistent with the other facts punched. Even these could he eliminated by comparison or check of every card. It is to he borne in mind, too, that a card wrongly punched involves only the possible miscounting of a single unit. whereas in all previous methods the counting up on sheets has involved possible miscount at each footing np of a column.
In the compilation of census statistics, such as those of population, mortality, etc., or the bulk of the work to which this apparatus has heretofore been applied, the person forms that unit, so that each card represents simply that unit. But the census includes agricultural, manu facturing and similar statistics, and it is evident that in Lhe figures of agriculture or maim facture, while a card might represent a farm or a factory unit, the value of that unit might vary greatly. Thus it might be a farm of 100 acres or of 500, and we would thus have to record amounts. This is done by a specially constructed machine containing a cylinder around whose circumference studs are set ; spring contact points connected to the mercury cups of the press ; a motor for revolving the cylinder, and a device for starting and stopping the motor so that the cylinder will make one revolution for each card. 'lime operation can be readily understood. A curd being put in the press. the eireuit is closed through a given counter to the battery, to the cylinder of the integrating device, from one of the nine con tact strips of the integrator through the corresponding mercury cup uncovered by the punched hole of the card through the plunger of the pin box corresponding to that hole, and back to the counter. At the same time, when the handle is brought down, another circuit is closed through the magnet, which allows the train to revolve the cylinder of the integrating device one revolution. During that revolution the circuit through the dial counter will be made and broken from one to nine times. according to the contact strip which is brought into operation. Any number of counters can thus be operated at the same time, they being con nected in multiple arc. The registration thus secured gives totals from any number of dif ferent sizes or amounts, and the device, therefore, answers a most useful purpose.
Tank, Glass : sec Glass-making, Tapering Machine : see Molding Machines, Wood.
Tapping. Machine : see Pipe-cutting and Nut-tapping Machines.