ROGERS, Circuit Judge
(after stating the facts as above). The question in this case brings before the court the validity of complainant’s patent, and, if the patent is valid, then the question of the defendant’s infringement.
It is the complainant’s contention that the patentee gave to the world the first wall radiator composed of units capable of being connected together to extend the area of radiating surface over a wall. The complainant insists that there is a substantial and effective difference between a floor radiator and a wall radiator, and that the features of construction and proportions of parts which render the patent in suit peculiarly effective, are patentable features and entitled to protection as such.
A radiator is described as:
“A device for delivering the heat, steam or hot water admitted to its interior to the air of the room or compartment which it is desired to heat.”
It is the function of a radiator to transmit the heat, which is supplied by the heating medium (steam or hot water) to the objects which are outside. It is the outside surface of the radiator which transmits the heat by contact with the air.
The art of warming buildings by radiators, located in rooms to be warmed, and connected by pipes with steam and hot water boilers, has been known to some extent for many years. In Hood’s Treatise on Warming Buildings, published in London in 1855, mention is made of [908]*908the fact that Count Rumford, who lived from 1753 to 1814, made improvements in steam heating apparatus.
The appellant complains that the court below erred in holding that claims 1 and 2 were not restricted by the specification and drawings of the patent in respect to the relative sizes of the outer cross, and connecting tubes,, of which the radiator sections are composed.
The opinion of the court below in respect to this particular matter is as follows:
“There was a prior suit brought upon this patent against a licensee who, of course, could not deny the validity of the patent. The licensee was dealing in a radiator, which was substantially in shape like the radiator described in the complainant’s patent, except that its various connecting tubes were substantially of about the same size. In that case, Judge Buffington originally held on a motion for a preliminary injunction, that the defendant’s radiator was substantially the same as the complainant’s radiator, but upon final hearing, he held that, in view of the prior art, the complainant’s invention was a very narrow one, and that the claims must be interpreted in the light of the specifications and drawings, which stated and showed that the top and bottom tubes were larger than the cross-tubes, and that the cross-tubes were larger than the small intermediate cross-tubes, and held that as the defendant’s radiator was not so constructed it did not infringe. Some, of the witnesses for the complainant in the case at bar lay stress on the fact that, as shown in the drawings and described in the specifications, the outside tubes are larger than any of the intermediate tubes, and the intermediate cross-tubes are larger than the tubes between the cross-tubes, and claim that such a construction is important in aid of the rapid diffusion of steam throughout all parts of the radiator, and the effective expulsion of air and water from the radiator. But the answer to all these assumptions based upon the different sizes of the tubes shown in the drawings and described in the specifications is that no claim is made in the patent of any invention in those respects. The first and second claims in the patent apply to a unitary hollow casting consisting of tubes, and if the claims are good they give a monopoly of the use of such a unitary structure with tubes of any size. The fact that the specifications describe a structure with tubes of different sizes, is, of course, immaterial. An invention described in specifications, but not claimed, is not protected by a patent. As the first two claims are drawn, in my opinion, they are anticipated by the patents cited and by the condition of the prior art.”
We cannot agree with the court in the construction placed on claims 1 and 2. In his specification Fowler set forth as his substantial advance in the art the fact that the cross-tubes of his radiator are of larger size than the connecting tubes, which run at right angles. In our opinion it is clear that these relative sizes are not mere preferential or illustrative constructions, but that they are the particular features which characterize the invention and are embodied in the claims in controversy. The patentee states in his specification:
“The sections are preferably two longitudinal top and bottom tubes B B', united by two end tubes O O’, which are connected by a series of smaller longitudinal tubes J>. I prefer to employ intermediate cross-tubes B connecting the tubes B B' and having the longitudinal tubes D connecting with them. In the construction shown there are two of these large intermediate cross-tubes B located between the end tubes O O and dividing the small longitudinal tubes D into three sets or series. The small longitudinal tubes D are located at sufficient distance apart to form intermediate openings or spaces for the circulation of air between them. Each section thus constructed is composed of a series of communicating tubes, forming a unitary rectangular hollow frame.”
[909]*909We think that the word “preferably” refers to the rectangular shape and not to the relatively “large intermediate cross-tubes” and the “small longitudinal” or connecting tubes. In this we find ourselves fully in accord with the position taken by the court in the Third Circuit, and we agree in thinking that the very fact that the two sets of intermediate tubes are called in the claims, one “cross-tubes” and in the other “connecting tubes,” shows a purpose to differentiate them, and that the only ground of differentiation is found in the figures and specification in which latter the cross-tubes are described as “large intermediate tubes,” and the connecting tubes as “small longitudinal tubes.” See Fowler & Wolfe Mfg. Co. v. National Radiator Co., 172 Fed. 661, 663, 97 C. C. A. 187.
There will be found in the specification that the patentee repeatedly states that the horizontal tubes B B' C C are large and the series of longitudinal tubes D are small. He also states that as tubes of the size of the small tubes D are liable to offer considerable friction to the steam or water traversing them, he prefers to shorten their length by the interposition of the intermediate transverse tubes If. But the statement in the specification which reads “1 prefer to construct the section A with the tubes arranged in the manner described,” refers, as we understand it, to a section which is longer horizontally than vertically.
[1] Wc cannot understand why the court below should have reached the conclusion that au invention described in the specification but not included in the “claims” cannot be protected by a patent. We regard the law as well established that the claims of a patent are to be construed in the light of the specification. It is quite true that the claims fix the extent of the protection furnished by the patent. McClain v. Ortmayer, 141 U. S. 419. 12 Sup. Ct. 76, 35 L. Ed. 800; Sutter v. Robinson, 119 U. S. 530, 7 Sup. Ct. 376, 30 L. Ed. 492; Burns v. Meyer, 100 U. S. 671, 25 L. Ed. 738. But it is equally true that the specification may be referred to for the purpose of limiting the claim although not available for expanding it. McClain v. Ortmayer, supra. Dey Time-Register Co. v. W. H. Bundy Recording Co. (C. C.) 169 Fed. 807, 813 (3909). In Howe Machine Co. v. National Needle Co., 134 U. S. 388, 10 Sup. Ct. 570, 33 L. Ed. 963, Mr. Chief Justice Fuller, writing the opinion of the court, said:
“Doubtless a claim is to be construed in connection with the explanation contained in the specification and it may be so drawn as in effect to make the specification an essential part of it; but, since the inventor must particularly specify and point out the part, improvement, or combination which he claims as Ms own invention or discovery, the specification and drawing's are usually looked at only for ihe purpose of better understanding the meaning of the claim, and certainly not for the purpose of changing it and making it different from what it is.”
We recognize the full force and effect of this statement of the law and do not believe that we are going contrary to its true meaning in holding that there can be read into the claims in this suit the qualification found in the specification that the tubes D are smaller than the others. By reading this qualification into the claims, they are made narrower and not broader.
[910]*910In Duncan v. Stockham, 204 Fed. 781, 123 C. C. A. 133 (1912) the court in the Seventh Circuit said:
“The claim in suit does not name all the various means shown in the specifications and drawings for connection of the means or elements named therein to make them operative in the combination; but we believe the claim is, nevertheless, sufficient for enforcement, on reference to the specifications, it is to be interpreted to include such connections and relations of the several means of the combination which are .named, as implied therewith to make them operative, in conformity with the specifications.”
In Consolidated Roller Mill Co. v. Walker, 138 U. S. 124, 11 Sup. Ct. 292, 34 L. Ed. 920 (1890), the Supreme Court declared it saw no reason to doubt the correctness of the following statement made by the court below:
“To understand the nature of the invention intended to be covered by the first claim, resort must be had to the specification, and we there find that the ‘swivel boxes’ are essential to the contemplated greater movement at one end of the shaft than at the other, whereby is effected ‘the tightening of the belt or belts at one side of the machine, without disturbing those at the other.’ This is apparent on the face of the paragraph hereinbefore quoted at length; and th¡e expert testimony is direct and convincing, that, to the practical working of the described device as a belt tightener, this swiveling feature is indispensable. Without the swiveled boxes Gray would not have ‘independently adjustable bearings.’ True, those boxes are not expressly mentioned in the claim; but we think they are to be regarded as entering therein by necessary implication, for the reason just stated, as well as by force of the words ‘as shown.’ Moreover, the prior state of the art would limit the claim to the specific organization shown and described.”
So in a number of cases the Supreme Court has sustained the validity of a patent which otherwise might have been invalid by importing into the claim the particulars of the specification. See Seymour v. Osborne, 11 Wall. 516, 547, 20 L. Ed. 33 (1870); The Corn Planter Patent, 23 Wall. 181, 218, 23 L. Ed. 161 (1874); Westinghouse v. Boyden Power Brake Co., 170 U. S. 537, 558, 18 Sup. Ct. 707, 42 L. Ed. 1136 (1897); Carnegie Steel Co. v. Cambria Iron Co., 185 U. S. 403, 432, 22 Sup. Ct. 698, 46 L. Ed. 968 (1901).
[2] The District Judge held that the cross-tubes B in the Fowler patent served no other function than that of radiating elements. In this view he agreed with the opinion of the primary examiner in the Patent Office. The testimony of the experts does not, however, in our opinion, support the conclusion. The professor óf experimental engineering in Cornell University, who testified that he had studied the underlying principles relating to the art of heating and had given special attention to the use and efficiency of radiators, was questioned concerning the following statements in the specification in the Fowler Eatent:
“The small tubes are perfectly made of an area substantially equal to that of a one-inch pipe; but as tubes of that size are liable to offer considerable friction to the steam or water traversing them, I prefer to shorten their length by the interposition of the intermediate transverse tubes U.”
His testimony in reply was as follows:
“It is a well-known' fact that if the friction is lessened in the flow of any fluid through a pipe, that the forces which act to move the fluid would pro-[911]*911¿Luce a greater velocity, or, in other words, thereby increase the circulation. This statement then in effect, explains how the cross-tubes EE increase circulation by reducing tbe friction. Tbe increase of circulation tends to remove the air from the radiator, which is usually found in considerable volumes in both water and steam and. which prevents, in a large measure, tlie attainment of a uniform temperature throughout the radiator. The obvious effect, then, of the improved circulation, is a uniform temperature throughout every portion of the radiator, and this condition the specifications indicate as in a measure due to the transverse or cross-tubes EE.”
Another witness, a graduate of Yale University, who had been for several years an assistant examiner in the United States Patent Office and was at the time of the trial a mechanical and electrical engineer, who was familiar with the subject of radiators and who had had wide experience, having been for ten years associated with one of the prominent companies engaged in the business of manufacturing and installing steam heating apparatus for railway cars, was also examined on the same subject. His testimony need, not be set out at length. It suffices to say that it was in entire accord with that of the witness to whose testimony wc have referred. There is nothing in the record that we have been able to find which contradicts the testimony of these witnesses. And irrespective of their testimony we should be inclined to the opinion which they have expressed. Tlie 'examiners in chief in the Patent Office thought that the cross-tubes had another function than that of mere radiating elements and that they possessed specific utilities and were clearly patentable. We think the testimony clearly sustains the conclusion. The primary functions of the cross-tubes are twofold: First, their action on internal circulation in facilitating the discharge of air and water of condensation; and, second, their fiction on external circulation in directing the air currents which pass over the radiating surface. They increase the efficiency of the radiating unit, not by the mere addition of radiating surface, but by aiding the internal circulation of the steam and the external circulation of the air.
The court below was of tlie opinion that the first two claims had been long anticipated in the prior art. A patentee’s claim to an invention is anticipated when it appears that another made the invention before the date when the patentee made it. And to authorize the allowance of a patent it is of course necessary that there should be a substantial difference in principle from prior inventions. But nothing can be regarded as an anticipation which is not a full and complete disclosure of tlie invention to the public such as will enable those skilled in the art to make and use it. It is well settled that to amount to anticipation, the two things accomplish the same purpose by substantially the same means operating in substantially the same way.
When the Fowler patent was before the court in the Third Circuit in Fowler & Wolfe Mfg. Co. v. National Radiator Co., 172 Fed. 661, 97 C. C. A. 187 (1909), it was not necessary to pass upon the validity of the patent. That suit, like this, was for an infringement, but the defendant was a licensee in the former suit and under tlie circumstances the validity of the patent was not one which could be properly raised. But in the present suit the validity of the patent is directly involved.
[912]*912The District Judge, in holding that the Fowler patent had been anticipated in the prior art, stated that the first two claims were for a structure which was shown substantially in the prior patents of Reed, Safford, Wood, and in the Bundy Pyro wall radiators. And that it was also suggested “in the somewhat analogous art of boiler making in the boiler used as a radiator in the Exeter Machine Works.”
In the Reed patent, No. 347,127, August 10, 1886, we are unable to discover the combination of parts which is found in the patent in suit. It is a radiator composed of sections of the loop form, having a central vertical tube connected with the outer tube by short hollow bosses. One of the defendant’s experts found in it an anticipation of claims 1 and 2 of the Fowler patent. He does not appear to be sustained in this conclusion by any other expert on either side. The Reed radiator has three vertical tubes, the middle one appearing to be somewhat larger than the other, the latter being about the same size as the header tube. But there is nothing in the Reed patent which shows, describes, or even suggests the intermediate cross-tubes and the series of connecting tubes which we find in the patent in suit; nor is there revealed a structure in which the parts would have the same functions and operate to produce the high efficiency attained in the Fowler patent. There is nothing to induce- perfect circulation and uniform temperature as in the patent in suit.
In the Safford patent, No. -355,216, December 28, 1886, we also fail to discover the combination of parts which the complainant claims. The Safford radiator, like the Fowler and Reed radiators, is composed of sections. It consists of a series of vertical radiating tubes. ' These vertical tubes are four in number and as the specification states are connected between the top and base portions by short horizontal tubes. The short horizontal tubes “form braces which strengthen the vertical radiating tubes,” and also form communications between the series of vertical tubes. There is nothing, as it seems to us, in the Safford specification or claims which suggests the group of tubes D in the Fowler patent. The Safford patent is for an ordinary form of floor radiator. The arrangement is materially different from that of the Fowler patent. In the latter the sections are. constructed with large outer or peripheral tubes and similar large cross-tubes to serve as the conducting and quick distributing means, together with the groups of small tubes set into the several panels formed by the larger ones and opening into the large ones at both ends.
In the Wood patent, No. 176,915, May 2, 1876, there is not shown a radiator having large outer tubes, a large cross-tube and a series of small connecting tubes. But the tubes are all of the same size. In the Fowler patent the relative dimensions of the tubes, as well as their number and arrangement, is essential to enable it to perform its functions and to accomplish its advantages. The three short connections of the Wood radiator are not at all the equivalent of the series of small connecting radiating tubes of the Fowler radiator. They are not designed for, neither are they capable of performing, the same functions. Wood states specifically that his radiator is constructed so as to allow for the free expansion and contraction of the different parts [913]*913of the radiator without impairing the joints. This is not in harmony with the idea of Fowler. As we have already stated the idea of the patent in suit is a construction such as to induce perfect circulation and uniform temperature. A construction of the Fowler radiator needs no allowance for the free expansion and contraction of the different parts-.
In the Bundy Pyro patent the radiator consists of a tubular section having headers at the top and bottom connected by a series of vertical tubes. There are solid brace pieces, or bosses, between the vertical tubes, but no cross-tubes or connection between them. There are no large outer vertical tubes with a cross-tube between opposite outer tubes and a series of connecting tubes between the cross-tube and the outer tubes. Moreover, the crosspieces between the vertical tubes are merely solid braces and can have no effect whatever upon the circulation. This fact alone is sufficient to show that a fundamental difference exists between the Bundy and the Fowler radiators. .It is perfectly obvious that the Bundy radiator does not embody the features of the Fowler radiator.
We pass on to a consideration of the Exeter sectional boiler to which the court below alluded in its opinion. It is a steam- boiler for generating steam composed of a series of cast iron sections. They have been manufactured under the Brayton patent, No. 53,399 granted in 1866. It is claimed that some of these boiler sections were used as wall radiators and tha.t they embody in all material respects the construction described in claims 1 and 2 of complainant’s patent. We are, however, unable to discover that these boiler sections embody the construction of Ihe claims of the Fowler patent, or that there is anything which in the slightest degree suggests the relative dimensions of the tubes which are necessary to produce the results attained under the latter patent. There are no large outer tubes and large cross-tubes B of the Fowler patent with a series of relatively small connecting tubes D, such as are essential in the Fowler radiator. If we were to attempt to consider the central cross portion as the cross-tube of the Fowler patent and the vertical portions as the connecting tubes, we should be constrained to abandon the idea by the fact that these parts are all of substantially the same dimensions. The functions and effects of the Exeter boiler section when used as a radiator are essentially different from the structure claimed in the patent in suit. In the Exeter section the cross-section of the space on the interior of the radiator between the vertical tubes was 2% inches by 2J1/io inches, and the cross-section of the space on the interior between the end of the vertical tubes was 2% by 2% inches. It is manifest that these dimensions as well as the section itself indicate the absence of the tubular construction which characterizes the structure of the patent in suit. The dimensions are not such as to induce circulation or to provide an efficient relation between heating surface and area. The absence of tubular construction in the Exeter section prevents circulation. There would be simply a body oí steam and possibly air over the water without any marked circulation of any character, and certainly none substantially similar to the circulation attained in the Fowler patent. We are wholly unable [914]*914to discover that the Exeter boiler in any way affects the novelty of the Fpwler construction.
The opinion of the District Judge contains no allusions to the Backus patents and perhaps it is unnecessary for us to allude to them. The Backus patents, however, were granted June 29, 1886. Patent No. 344,512 provides for a device intended for generating steam by the heat of gas in a radiator placed in a fire place. Patent No. 344,511 relates to a combined heating and cooking device. As described in the specifications:
“The object of the invention is to provide a household article which will be valuable as a heating-stove, a cooking stove, a boiler and steam radiator, and an ornament for a room.”
The radiator of the Backus patent Ño. 344,512 does not have the same kind of circulation as the radiator of the patent in suit. And it does not have the same structure as there are no large cross-tubes connecting the other tubes. It operates in a different manner and does not produce the same result. And what has been said of No. 344,512 may be said also of No. 344,511. The latter does not differ in type from the former It would unduly extend the limits of this opinion for us to refer in detail to each of the 23 prior patents to which the attention of the court below was invited and which we find fully described in the record. An examination into these various devices has failed to convince us that claims 1 and 2 of the patent in suit are invalid because of anything to be found-in the prior art.
An examination of the prior art as disclosed in the record has however convinced us, as the court in the Third Circuit was convinced when this patent was before it in the case before mentioned, that the invention of the patent in suit is a narrow one-.
Claims 3 and 4 stand upon a different footing from claims 1 and 2 .and we must hold them to be void. They merely provide old means for coupling the sections when they are laid together. Everything in claims 3 and 4 except the coupling devices, including the holes as part of the coupling device, is shown in claims 1 and 2, for it is manifest that the “unitary radiator section” is susceptible of indefinite multiplication. The single new feature contained in claims 3 arid 4, that of coupling the sections, is not patentable invention.
We pass now to consider whether the proofs show that the patent in suit has been infringed by the defendant. The bill alleges that the defendant against the will of the complainant and in violation of its rights has infringed upon letters patent 609,800 “by selling, since the date of the said letters patent, radiators substantially the same in construction and operation as in the said letters patent are described, and is now so engaged in offering for sale and selling the same * * * to the great gain and profit of the defendant and to the great loss and injury of your orator.”
The proofs establish the fact to our satisfaction that the defendant’s radiator possesses the same functions and effects which the complainant’s radiator possesses and which are specified in claims 1 and 2 of the patent in suit. We insert a drawing illustrating the defendant’s radiator.
[915]*915If wc compare the above drawing with the drawing herein before inserted of the patent in suit, it is quite evident we think that the outer tubes of the defendant’s radiator are B B' C C of the patent in suit; that the tubular passageway cut through the intersecting tubes and made steam tight by the tubular bosses is cross-tube B of the patent in suit; that the remaining tubes are concededly radiating tubes, arranged in panels in series, and, since it does not seem to be disputed that they are smaller than B B' C C and B, they are the tubes D of the patent.
In conclusion it may be remarked that the defendant’s own conduct seems to indicate that it is fully aware of the meritorious character of the complainant’s invention in making these panel series of small diameter. If there be no merit in these relative proportions, as defendant contends, it impresses tis as strange that defendant insists on making its series of radiating tubes of smaller diameter than the B, C, etc., of the patent.
The decree is reversed, with the direction to grant a perpetual injunction as prayed, and that an accounting be had as to claims 1 and 2 of the patent in suit which we hold infringed by the defendant company, and that the complainant recover its costs.