section thus constructed is composed of a series of communicating tubes, forming a unitary rectangular hollow frame.' Now it is clear to us 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, No other sizes or modes of construction are shown or suggested and the larger sizes are designated as 'cross-tubes' and the smaller as 'connecting tubes,' and as such, both are carried into the claims. The very fact that the two sets of intermediate tubes are called in the claims, one 'cross-tubes' and the other 'connecting tubes', shows a purpose to differentiate them and 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'. These limitations, if carried into the claims, embody a different structure from that of the respondent and the claims so read fail to sustain the charge of infringement. So construing the claims infringement is not shown and the bill will be dismissed."

After a careful consideration of the record and arguments of counsel, we are of opinion that the court below were justified in the views thus expressed.

An examination of the prior art, as disclosed in the record, shows at once that the invention of the patent in suit is a narrow one. Arrangements of cast and wrought iron radiating tubes placed in juxtaposition, in sections, and parallel with each other, terminating and opening into larger tubes at the top and bottom, thus affording an uninterrupted circulation of steam or water through all of them, were quite old in the art. They constituted unitary hollow quadrilaterals or castings (where cast iron was used instead of wrought iron, though that is not material), consisting of four outer tubes communicating with each other at the corners, with a series of transverse and parallel radiating tubes, extending between two of the outer tubes and opening into them. Examples of such structures are shown in the Mills radiator patent, No. 154,651, the Safford radiator patent No. 355,216, and others, constituting in their general features a standard radiator structure in use long before the date of the patent in suit. In view of this standard form, the device of the patent in suit exhibits only a somewhat different arrangement of tubes and a division of the cast iron section into sub-sections, by what are called cross-tubes in the specifications and claims. The words "cross-tubes" convey no definite and adequate meaning, when taken alone. They necessarily are relative. The words of the claim are, "one or more intermediate cross-tubes between opposite outer tubes." They must be taken relatively, and therefore to require outer tubes, opposite to each other, between which they must continuously extend as unitary structures, and they must correspond structurally and in fact to what is connoted by the word "tube." We look at once at the drawings and specifications of the patent, as we are authorized to do in order to obtain a more definite conception of these cross-tubes relatively to the other parts of the structure, and we find in the drawing of the patent the two cross-tubes extending between the longer outer tubes of the section and corresponding in size therewith. This drawing also makes clear what is meant by a series of connecting tubes between the cross-tube and the outer tube, or between two cross-tubes, which was not clear from the claims themselves. When we turn to the specifications, the character and relation to each other of the tubes constituting the unitary hollow casting dealt with in the

claims, becomes entirely clear. We quote this part of the specifications, and insert a drawing taken from the patent in suit, illustrating complainant's radiator:

"A is one of the radiator sections, composed of a hollow cast-iron frame made up of a battery of tubes.

"The sections are preferably rectangular in shape and composed of two longitudinal top and bottom tubes BB', united by two end tubes CC', which are connected by a series of smaller longitudinal tubes D. I prefer to employ intermediate cross-tubes E, connecting the tubes BB' and having the longitudinal tubes D connecting with them. In the construction shown there are two of these large intermediate cross-tubes E located between the end tubes CC 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 e 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.

"I prefer to construct the section A with the tubes arranged in the manner described, because as the circulation of the steam and water is in horizontal planes the most free circulation should be permitted in those directions, and this is attained by the large horizontal tubes BB' and the series of small tubes D. The small tubes are preferably 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 E. While I prefer this particular arrangement of the tubes for the reasons stated, it is to be understood, however, that I do not mean to limit myself to this precise construction."

The frequent use of the word "preferably" throughout this part of the specifications, does not serve to extend the scope of the claims beyond the structure thus shown, where no alternative form is suggested and the relative sizes of the cross-tubes and connecting tubes shown in the specifications are essential to their respective functions. But, aside from this, it is noticeable that the words "preferably" and "I prefer" are immediately applicable to the rectangular shape of the section,

or to the mere arrangement of the tubes, as being horizontal or vertical. The "cross-tubes" are throughout referred to as larger, and the connecting tubes as smaller, which the latter must necessarily be, as there is a multiplicity of them, as indicated by the word "series." Indeed the presence and use of cross-tubes is due to the necessity of shortening. the series of connecting tubes, which, by reason of their small diameter, would offer too much friction if they extended across the rectangle. They were therefore shortened by the interposition of the intermediate tubes. As these intermediate cross-tubes take the place of the large outer tubes, as features of the multiplicity of connecting tubes, they must of necessity be relatively larger. It is thus seen that, functionally this limitation is imposed upon the language of the claims. On this point, Mr. Bently, the expert witness for the defendant, says:

"Thus it seems but reasonable to understand that when Mr. Fowler refers In his claims to the cross-tubes 'between opposite outer tubes,' he means such cross-tubes as he describes and illustrates, to wit, tubes of substantial size and capacity, capable of supplying the small connecting tubes from the outer tubes at one or more intermediate points along the length of the radiator sections. Moreover, when he speaks in his claims of 'a series of connecting tubes,' he means such connecting tubes as he shows and describes, to wit, a substantial multiplicity of small tubes, having relatively large radiating surface, but short and having their conducting capacity reinforced by the crosstubes."

It is true, as the appellant contends, that the claims of the patent are not to be limited beyond the clear import of their terms, but it is equally true that we may refer to the specifications and drawings to discover the meaning of the terms, as used in the claims. The terms "outer tubes," "cross-tubes" and "connecting tubes," are not technical terms recognized and employed in the art. They are terms used by the patentee to apply to a peculiar construction and arrangement of tubes unknown before his invention. "To ascertain the meaning of these terms, we must therefore go to the specifications, not for the purpose of importing anything into the claims, but for the purpose of ascertaining what characteristics are attributed in the specifications to the terms therein contained." Consolidated Co. v. Walker, 138 U. S. 132, 11 Sup. Ct. 292, 34 L. Ed. 920; Turril v. Railroad Co., 1 Wall. 511, 11 L. Ed. 668; Brooks et al. v. Fiske et al., 15 How. 215, 14 L. Ed. 665. To these authorities, cited in appellee's brief, many others might be added. The discussion of the claims of a patent is not to be confined to a mere logomachy, which could be set at rest at once by a reference to specifications and drawings. The radiator of the defendant is the old standard radiator of the prior art. That is, there are two horizontal headers united by a series of vertical radiating tubes, through which the steam or hot water circulates from header to header. Without more, it is not contended that such a structure infringes the patent in suit.

But it is contended by the appellant that the distinguishing feature of the radiator of the patent in suit, to wit, a cross-tube, is present in defendant's structure, inasmuch, as, halfway the length of the vertical tubes, and between them, are hollow bosses, which form a passageway for the steam through the entire series of vertical tubes. Here

again, it is necessary to inspect complainant's exhibit of defendant's structure, in order to understand the meaning of this contention:

END OPENINGS

It would seem perfectly apparent from such an inspection, that there is nothing here to answer to the description of the cross-tube in plaintiff's claim, as explained in the specifications. Relying on the literal words of the claim, these connecting bosses do not singly or together constitute a "cross-tube between opposite outer tubes." They may constitute a passageway from one connecting tube to the other, and in connection with these tubes themselves form a passageway from one outer tube to the other, but this passageway is not a tube, nor is it created by a tube. Again, if this series of links in the passageway from one outer tube to the other could, by any ingenuity be interpreted as a cross-tube, it is evidently much smaller than the connecting tubes, instead of larger, as we have shown the cross-tubes of the patent in suit must be. It is sufficient to say, however, that the passageway thus created is not a cross-tube, or a tube of any sort, and neither literally nor functionally comes within the scope of the claims of the patent in suit. Again, if the passageway made possible by these small hollow bosses constitutes a cross-tube, then the vertical tubes of defendant's radiator are the connecting tubes "between such cross-tube and the opposite outer tube," but it is plain that these connecting bosses do not divide or shorten these so-called connecting tubes, which are integral between the headers and constitute structurally continuous tubes from one header to the other. It is an almost absolute reversal of the arrangement of the tubes of the patent in suit. It is curious to note that

RADIATING

RADIATING

TUBE

PRIOR ART

the appellant's expert in the license suit, when discussing defendant's radiator, designated the vertical tubes between the outer headers of defendant's structure, as "cross-tubes," and the series of hollow connecting bosses as "connecting tubes," while the expert in the present suit, as we have seen, calls these vertical tubes extending from header to header, "connecting tubes," and the series of hollow bosses between them, "cross-tubes."

That the claims must be read in the light thrown upon the relative size of the outer cross and connecting tubes by the specifications, is confirmed by the significant statement made by the appellant company in its catalogue published after the issuance of the patent in suit. On the first page of this catalogue, under a picture of the patented radiator, representing the outer tubes and cross-tubes as large and the series of connecting tubes as small, there is the following statement:

"The large, outer and cross-tubes, connecting and intersecting the smaller, insure the most positive circulation, the quickest condensation and the most rapid heating under all conditions. These essential factors are covered by our patents. The dimensions insure the entire surface to be effective and efficient."

It is hardly necessary to say that the defendant's radiator presents none of these "essential features." In fact, the defendant's radiator is the radiator of the prior art. Of the "Safford" patent, No. 355,216, December 28, 1886, the expert witness for the defendant says this patent "shows a cast-iron radiator section of the type in question, having four parallel tubes connected across the top and bottom by headers, and moreover, the parallel tubes are in communication with one another by a perforated boss, just as in the defendant's radiator." The same witness also further testifies:

"Reed patent No. 347,127, August 10, 1886, shows another example of the defendant's type of radiator, comprising a number of parallel tubes connected across at their ends by headers. The Reed radiator is also a cast iron radiator. Moreover, it has the same incidental feature of perforated bosses which is found in the defendant's radiator, permitting communication between the parallel tubes."

A reference to these patents and the accompanying drawings, as shown in the record, fully confirms the testimony of defendant's expert. There was also the testimony of more than one witness qualified by foundry experience in the manufacture of cast iron radiators, to the effect that these connecting bosses were used solely as stays between the long parallel tubes. The result of this testimony is thus 'given by defendant's expert:

"The explanation for their presence is, that in casting these radiators, it is necessary to use what is known as a 'core,' which is a fragile rod of sand or similar material, placed in the mould to form the interior bore of the tube. When these cores are of small diameter and of considerable length, it becomes desirable to brace them, one against the other, by a small cross-core, and this cross-core, when removed, leaves a small opening from one of the parallel tubes into the next. It is obvious to any mechanic, that the presence of this opening between the adjacent parallel tubes of the defendant radiator, is merely an incident of the casting operation."

It is hardly necessary, after the foregoing statement of the interpretation to be given to the language of claims 1 and 2, and of the