i8S 



NATURE 



[April 24, 19 13 



I will now consider in various actual examples the 

 nature of \j/, the inventive function, and of i, the 

 inventive increment. 



One of the commonest cases in which a decision is 

 necessary is that of a combination. Suppose that 

 f(a) and j(b) are old ; will there be invention in com- 

 bining a and 6 ? 



The answer is this : 



(1) l=/(a,/>) 



(2) M =/(«)+/(*) 



(3) I = M+*. 



If the result of the combination is given by (i), 

 there is an invention; this is termed a "combination." 

 If the result is given by (2), there is no invention; 

 this is termed an "aggregation." It is interesting 

 to compare this definition with one given by Lord 

 Justice Buckley (Brit. United Shoe Mach. Co. v. 

 Fussell) of a " combination " as " a collocation of inter- 

 1 ommunicating parts, with a view to obtaining a 

 simple result." 



An example of a true "combination" is found in 

 the case of Cannington v. Nuttall, in which a patent 

 was upheld for a glass furnace, although each and 

 every part (a, b, c) had been employed before in glass 

 furnaces (employment = /). But, owing to the com- 

 bination, and the co-operation of the parts, a new 

 result was obtained. 



l=f(a,6,c)=f{a)+f(b)+f{c)+i. 



On the other hand Bridge's case is an example of 

 an aggregation; in fact, a patent was refused by the 

 Law Officer, showing that the case was considered 

 absolutely devoid of invention. The alleged invention 

 consisted in the employment in a shutter for dividing- 

 up rooms (/) of means (a) to guide the shutters along 

 the floor, and cogs (6) to hold the shutters against 

 the wall, /(a) and f(b) were both old, and no new 

 result flowed from their juxtaposition. Hence 

 M=/(a) + /(6): there was no invention; each part 

 simply played its own r61e, and there was no inter- 

 action. 



Another type of invention is that of varying propor- 

 tions in a known combination. Here, if M=/(a, b, c), 

 and if there is a maximum at one value or range of 

 values of c, invention may be involved. The maxi- 

 mum relates to the technical effect, and may be with 

 respect to efficiency, economy, &c. 



Thus if ""' ' c <> = o at the value c,, the function will 

 dc ' 



be a maximum or a minimum and there may be an 



invention. This will not be the case if Zy 7 '- ' f " 4= o. 



dc 

 Other singular points may be inventions, e.g. where 



3/(<yW = 00 (discontinuity), or where B ^ a ' f' c ^= 00 



(kink in the curve). This also holds for a range of 

 values from c, to c 2 . 



Examples of the application of this equation are to 

 be found in the cases of Edison v. Woodhouse, and 

 Jandus Arc Lamp Co. v. Arc Lamp Co. In Edison's case 

 / represented the employment in an incandescent lamp 

 of an exhausted glass vessel (a), leading-in wires (6), 

 and a carbon filament (<:,). /(a, b, c) was known, 

 but it had never been proposed to use a very thin 

 carbon conductor or " filament." Here, owing to the 

 high resistance and flexibility of the filament, the 

 -Hi. icncy was a maximum : — 



"' - -1 =0, and the choice of this value <,, which 

 dc 



made the difference between failure and success, was 

 held to be an invention. 



In the Jandus Arc Lamp case, / represented the 

 employment in an arc lamp of carbons (a), a tightly 



NO. 2269, VOL. 91] 



= when f] = 3 in. 



fitting sleeve (b), and an envelope of glass, &c. (c), 

 inside the outer globe. By making the glass envelope 

 3 in. in diameter a maximum efficiency was obtained, 

 and on this ground the patent was upheld, although 

 envelopes had previously been made 9 in. in diameter. 

 Here again : 



Bf(a,J,.c,)_ 

 dc 



A further example is an old case (Muntz v. Foster) 

 in which a sheathing for ships was made of sixty 

 parts of copper (a) and forty of zinc (b). 



Alloys of copper and zinc had been used before in 

 about the same proportions, but in this case the same 

 result would not have been attained, because Muntz 

 specified the best selected copper and highly purified 

 zinc. The impurities (Sx) were of great and unsus- 

 pected importance. Moreover, other alloys of copper 

 and zinc (probably even of purified metals) had been 

 made. We may consider the two points separately. 



(1) Impurities : — 



f(a + Sx, b + Sx) was old, where Sx represents impuri- 

 ties. Muntz's alloy was f(a, b) = f(a + Sx, b + Sx) + i, 

 hence there was an invention. 



(2) Selection of 60 : 40 percentage : — 



J\" _so > 'in) _ Q s | nce a t this percentage the efficiency 

 dd 

 was a maximum, because the alloy oxidised just fast 

 enough to prevent barnacles adhering to 'he ship, but 

 not fast enough to waste away excessively. 



On the contrary, the case of Savage v. Harris was 

 one in which there was held to be no invention in 

 changing the size of part of a device for retaining 

 ladies' hats in place. There was a back portion (a) 

 and teeth (6), and the size of the back was altered : — 



■'\f" — -4=o, and there was no invention. 



A known device or material (a) may be employed 

 for a new purpose ($). If /(a) is the old use, and 

 #(a) the new use, we have for an invention 

 \ = 4>(a) = f(a) + i. But if M= </>(a)=/(a), there is no 

 invention. The oft-quoted case of Harwood v. Great 

 Northern Railway Company was one of the latter 

 type. Fishplates (a) had been used for connecting (/) 

 logs of timber, and it was held there was no invention 

 in applying them (<t>) to rails in which they acted in 

 the same manner : — 



*(a)=/(a). 



But in Penn v. Bibby, wood (a) was employed (0) 

 for the bearings of propellers in order to allow the 

 water to pass round the friction surfaces. Wood had 

 previously been employed (/) in water-wheels, but 

 <£(a) = /(a) + i, and it was held that there was inven- 

 tion. 



A similar type of invention is that in which different 

 materials are employed in the same process. Here 

 f(a) is old, and f(x) is new. If f(x) = f(a) there is no 

 invention. If f(x)=f(a) + i there is invention. In the 

 recent case, Osram Lamp Works v. Z Lamp Works, 

 a patent was upheld for the use (/) in incandescent 

 filament lamps of tungsten (x), though osmium (a) 

 was known. Tungsten was more efficient and 

 cheaper : — 



j(x) = f{a) + Si, where Si represents a small degree 

 of invention. This in itself might not have been 

 sufficient, but it was coupled with the fact that one 

 particular process of removing the carbon from the 

 filaments was selected out of three known processes. 

 This may be considered to require an amount of 

 ingenuity Ai". Si+Ai = i, and therefore f(x)—f(a) + i, 

 and there is invention involved. 



Another type is the omission of one step in a known 

 process. In the case of Badische Anilin- und Soda- 

 Fabrik v. Soc. Chim. des Usines du Rh6ne, it was 



