*1202 FAIRCHILD, Chief Judge.

Plaintiffs in this action appeal from the district court judgment 1 of invalidity of Patents No. 3,241,876 (’876) (application filed January 31, 1964 and patent issued March 22, 1966) and 3,309,131 (’131) (application filed February 11, 1966 and patent issued March 14, 1967) and of non-infringement by defendants of the patents. 2

W. Selden Saunders (Saunders) is the patentee and owner of patents ’876 and ’131. Rudkin-Wiley Corporation (Rudkin-Wiley) is the exclusive licensee under the two patents. The patented devices are designed to reduce wind resistance in, respectively, a tractor-trailer combination and a single chassis truck. Saunders and Rudkin-Wiley on April 8, 1974 instituted this action for infringement against defendants Air-Flo Company (Air-Flo); Robert G. Geiger (Geiger), President of Air-Flo; and Recreational Supply and Equipment. The district court found the patents invalid as being anticipated by and obvious in light of the prior art. 3 The district court also found that defendants’ product had not infringed the Saunders patents. 4

The Saunders device is a baffle mounted above the roof of the cab of a tractor of a semi-trailer in the case of ’876 5 and of a truck in the case of ’131. 6 The purpose is to *1203 divert the air flow so that it will not strike the front of the trailer or truck body and will reattach smoothly to the top of the trailer or body. In the case of the semitrailer the diversion also avoids entry of the air stream into the gap between tractor and trailer, and a low pressure air eddy occurs in the gap. The advantage is that the air drag is less than would occur if the air flow struck the face of the trailer without being deflected by the baffle. Fuel consumption is reduced.

As stated in the application for ’876, ‘the device of the instant invention produces a relatively wide diffusion of the air impacting the forward portion of the trailer, and causes the same to readhere to the body of the [trailer] rearwardly of the front portion thereof in a relatively smooth and even manner, while at the same time creating a low pressure area or bubble between the tractor and the front of the trailer, so that the trailer will, in effect, be pushing forward against reduced rather than increased resistance.’ Claims 1 and 2 of ’876 position the baffle

forward of the face of the trailer at a distance approximately 0.7 of the half-width of the trailer. The baffle is inclined rearwardly, and forwardly convexed in a horizontal plane. Its height is substantially 0.7 of the difference in height between the tractor roof and the trailer. Most of the claims of ’131 position the baffle at the top of the windshield and do not state a formula for height.

35 U.S.C. § 102 contains the requirement of novelty for a patent. The district court found that “[t]he respective combinations and arrangements of parts shown in the Stamm patent and in the University of Maryland (UM) publications are the same as those claimed in the Saunders patents, and hence are effective anticipations under 35 U.S.C. § 102(a) .435 F.Supp. at 300. The district court erred in finding anticipation. Its key error in this regard was its view that because the Saunders baffle, the UM fairing and the Stamm conduit each can be said to deflect air, the fairing and conduit anticipated the Saunders patents. As this court has indicated, “[a] previous patent . .. anticipates a purported invention only where, except for insubstantial differences, it contains all of the same elements operating in the same fashion to perform an identical function.” Popeil Brothers, Inc. v. Schick Electric, Inc., 494 F.2d 162, 164 (7th Cir. 1974). Accord, Reynolds Metals Co. v. Aluminum Co. of America, 609 F.2d 1218, 1220 (7th Cir. 1979), cert. denied, 446 U.S. 989, 100 S.Ct. 2976, 64 L.Ed.2d 847 (1980); Illinois Tool Works, Inc. v. Sweetheart Plastics, Inc., 436 F.2d 1180, 1182-83 (7th Cir.), cert. dismissed, 403 U.S. 942, 91 S.Ct. 2270, 29 L.Ed.2d 722 (1971). The differences between Stamm and UM, on one hand, and Saunders on the other, are *1204 substantial and are shown infra in the discussion regarding obviousness.

A patent is invalid under 35 U.S.C. § 103 for obviousness “if the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains.” 35 U.S.C. § 103. The district court considered the Stamm patent, the University of Maryland tests, and several bug deflector patents and other known principles in determining that the Saunders patents were invalid for obviousness.

As the district court recognized, the analytical steps for determining obviousness are set forth in Graham v. John Deere Co., 383 U.S. 1, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966).

While the ultimate question of patent validity is one of law, A. & P. Tea Co. v. Supermarket Corp., [340 U.S. 147 (1950)] supra, at 155, [71 S.Ct. 129, at 131, 95 L.Ed. 162] the § 103 condition, which is but one of three conditions, each of which must be satisfied, lends itself to several basic factual inquiries. Under § 103, the scope and content of the prior art are to be determined; differences between the prior art and the claims at issue are to be ascertained; and the level of ordinary skill in the pertinent art resolved. Against this background, the obviousness or nonobviousness of the subject matter is determined. Such secondary considerations as commercial success, long felt but unresolved needs, failure of others, etc., might be utilized to give light to the circumstances surrounding the origin of the subject matter sought to be patented. As indicia of obviousness or nonobviousness, these inquiries may have relevancy.

383 U.S. at 17-18, 86 S.Ct. at 693-694.

SCOPE AND CONTENT OF THE PRIOR ART

Crucial to the question of validity of the Saunders patents is an understanding of the aerodynamic devices under consideration: fairings, conduits, and baffles. Aerodynamicists have long recognized that reduction of aerodynamic drag (air drag) of a vehicle will increase efficiency and save fuel. We must now view the scope and content of the prior art considered by the district court.

The University of Maryland conducted over 7,000 wind tunnel tests of Trailmobile trailers in 1953 to determine whether various devices or design changes would reduce air drag. The tests were of “trailer bodies with rounded front corners, rounded rear corners, various tractor-trailer gaps, streamlined tail fairing for trailer, deflecting vanes and the removal of underneath accessories and cross members.” The tests were all concerned with methods of streamlining, so as to keep the air passing the trailer attached to it in order to avoid areas of air turbulence and reduce air drag. One of the methods used was attaching a full fairing 7 from the tractor to the trailer so that the air hitting the tractor cab and fairing would remain attached and flow along the fairing until it reached the top front of the trailer, from which point the air would then proceed along the top of the trailer to the rear, remaining attached. The fairing eliminated the gap between the tractor and the trailer, enclosing it entirely. The UM study recognized that such a fairing would be impractical because it would prevent the tractor from turning independently of the trailer, but concluded that “a fairing approaching this type that does not attach to the trailer may not be unreasonable.”

The Stamm patent, No. 2,863,695, applied for in 1954, issued December 9, 1958, discloses that Stamm considered the conclusions reached after comprehensive wind tunnel tests (evidently the UM tests) including the reduction of drag which would result from streamlining, and the infeasibility of enclosing the space between the tractor *1205 and trailer. He directed his attention to providing a device which would control the air flow between the tractor and trailer so as to significantly reduce the drag losses that normally result from the air flow turbulence in this region. His solution was a conduit type means on the rear top and side portions of the tractor cab, receiving the airstream passing over the tractor cab and guiding and discharging the airstream toward the top and side portions of the front end of the trailer. Alternatively he would, in addition, provide conduit means on the surface portions of the trailer to receive the airstream after it left the conduit on the tractor and discharge it along the surface of the trailer. As Stamm described one embodiment, in part:

“the cab conduit or shell member will cause airstream flow along the top and sides of the cab to be directed rearwardly in such a fashion that it will be directed into the trailer-mounted conduit shell member and thereafter discharged along the top and sides of the trailer body in a fashion that tends to reduce airstream turblence and thereby reduce drag losses.”

(References to drawings omitted.)

There is no evidence that the Stamm device was ever effectively used.

The Examiner cited as references in ’876 Stamm, Huet U.S. No. 2,234,906 (1941), and Rix Great Britain No. 734,735 (1955). A summary of the conclusions of the University of Maryland tests was set forth in Stamm.

Huet in part discloses deflecting devices adjacent the side edges of a car of a train at its rear end to deflect air currents outwardly away from the gap between that car and the car behind it.

Rix described other bug deflectors mounted at the front of an automobile hood, and disclosed as its subject an air deflector across the hood at a position nearer to the windshield than to the front:

‘Thus the air flowing over the bonnet top, instead of striking the steeply inclined windscreen, is deflected by the less steeply inclined deflector on the bonnet top and directed upwardly to or over the top of the windscreen, the streamlining effect thereby being enhanced, wind noise being replaced, the windscreen being kept clear for longer periods, and the driver’s view of the road not being interfered with in any way.’

A reference cited in ’131 was Herrmann, German No. 1,022,481 (1955), directed to an air deflector projecting above the roof of an automobile having a roof opening in a sliding closure.

The additional prior art considered by the district court, included the following:

a. A group of bug (or rain) deflector patents—Hoag, No. 2,206,956 (1940); Heintz, No. 2,220,715 (1940); Metzger, No. 2,229,516 (1941); Parke, No. 2,338,199 (1944); McVicker, No. 2,644,716 (1953) and French Patent No. 1,067,101 (1954). The district court said, “The patents for bug deflectors disclose the provision of a deflector on a front lower portion of a vehicle serving to separate air therefrom and avoid impact of the air flow on the front surface of a higher rear portion of the vehicle, such as a windshield.” Finding of Fact at 42. Rix, which was cited by the Examiner, is similar to this group.

b. Lillie, No. 1,777,569 (1930) disclosing an auxiliary windshield, the equivalent of the no-draft panels long used on automobiles.

c. Hornke, No. 2,914,231 (1959) relating to automobile luggage carriers adapted for roof mounting. The patent recognized that luggage carried atop a car created several problems, including an increase in wind resistance with a consequent lessening of gasoline economy. Hornke proposed use of an angled shield in front of the luggage carrier which would “reduce wind resistance against the carrier and protect the luggage against wind, insects, rain, and other foreign matter.” Hornke referred to the function accomplished by the shield as that of “provid[ing] streamlining to the carrier to minimize wind resistance .. .. ” The Hornke shield is directly attached to the luggage carrier at the front of the carrier *1206 and the drawings accompanying the patent show the shield to extend above the height of the carrier. A drawing depicting a cover for the carrier indicates that Hornke recognized that the top of the luggage might be the same height as the shield.

The district court found that the principle of and use of a windshield or baffle as a protection against air flow pressure and the creation of low pressure eddies behind a wind baffle was known and illustrated by the noted Italian scientist Leonardo da Vinci and was discussed in Fluid Dynamic Drag by Hoerner, before Saunders. The illustration in the Hoerner text referred to, however, was a tracing described as “Flow pattern past a bluff obstacle.” The text explains that the original drawing was “of flow patterns as [da Vinci] had determined them in water.”

The district court also found that “[i]t was commonly known by aerodynamicists that the spacing between a pair of bodies or the size of one body relative to the other will vary the drag on the second body as shown by experiments done in Paris in 1914.” The finding is supported by the testimony of defendants’ expert.

DIFFERENCES BETWEEN THE PRIOR ART AND THE CLAIMS

The patents and publications which were before the court disclosed that (1) streamlining of a tractor-trailer combination would lower air drag and improve gasoline efficiency; (2) a fairing connecting a tractor and a trailer would accomplish streamlining, but was not practical because it would not allow the tractor to turn; (3) a low pressure area would be created behind a windshield or other barrier preventing the passage of wind; (4) a baffle would cause the separation of air from a surface; and (5) Stamm suggested that conduits could be constructed on top of the tractor to divert the airstream so that it would not strike the face of the trailer, but travel along its top and side surfaces. There was testimony that the Stamm structure is not a baffle.

The University of Maryland test was concerned with streamlining, maintaining a flow attached to the vehicle with limitations on the turbulence of the air flow in the vicinity of the vehicle. The district court viewed the fairing in these tests as anticipating the Saunders baffle. 435 F.Supp. at 300. The problem we have with this conclusion is that the district court viewed “fairing” and “deflector” as interchangeable terms, 8 whereas the fairing does not deflect air from the surface but rather provides a new surface along which the air will flow. Plaintiff’s expert witness, aerodynamicist A.M.O. Smith, testified that the University of Maryland fairing was a streamlining device and did not create air separation.

A dictionary definition of “baffle” is “something for deflecting, checking, or otherwise regulating flow, as: A. a plate or wall for deflecting gases or other fluids . . .. ” Webster’s New International Dictionary, 3d ed. at 162. Witness Smith defined baffle as “something that is essentially broadside on, so it has separation behind it, ... ” and as “something that tends to obstruct the flow in some manner.” Smith contrasted baffle with conduit, which he indicated is “something through which a fluid flows ...” and as definitely not a baffle. The Stamm patent utilized a conduit through which the flow of air was channeled. In contrast, the Saunders baffle obstructed rather than channeled the flow of air. Although both can be said to deflect the flow of air, there are very significant differences in many respects, such as abruptness, turbulence, and resulting low pressure area.

In addition to achieving air deflection by a different method from Stamm’s, the Saunders baffle is structurally different and some of the claims call for a range of placement and size for the baffle which distinguish it further from either the University of Maryland or the Stamm prior art.

*1207 The various bug deflectors in the prior art are all baffles. Plaintiffs adduced testimony, however, indicating that bug deflectors on automobiles increase air drag. We cannot find any evidence to the contrary to support the district court’s finding that bug deflectors reduced air drag.

It is clear that prior art showed that air pressure would be reduced behind a baffle mounted on a moving vehicle. Windshields on open vehicles are illustrations. The Saunders’ claims indicate that the deflection of air will both create a low pressure area in front of the trailer and will divert air which will reattach itself to the trailer roof instead of striking the face of the trailer.

The size and positioning of the Saunders shield differ significantly from the size and positioning of the bug deflectors and the Hornke shield. In order to avoid interference with the driver’s vision, the bug deflectors were required to be much shorter in relation to windshields than the Saunders shield is designed to be in relation to the height from the top of the tractor to the top of the trailer. On the contrary, the Hornke shield is shown in the drawings to be significantly higher than the luggage rack and the same height as the luggage it is meant to shield.

Put in other terms, the prior art included the University of Maryland concept of altering the shape of the tractor cab by adding a fairing in order to reduce air drag by streamlining. It included the Stamm concept of a structure atop the cab at its rear to channel and turn the airstream so as to flow along the trailer without striking its face. It included the bug deflector concept that relatively small baffles on an automobile hood would direct air currents (and bugs) up and over the windshield. There is a known principle that varying the relative size of two bodies in tandem, and the spacing between, would vary the wind resistance of the combination. A further principle is that low pressure eddies exist behind a wind baffle.

The University of Maryland and Stamm devices differed from Saunders in that they are not baffles. Whether or not the bug deflectors reduced air drag of an automobile, they were designed for other objectives. Although the known principles referred to are consistent with the reduction of air drag produced by Saunders, there is no evidence that anyone previously sought to reduce the air. drag of a tractor-trailer combination by placing a baffle atop a tractor cab, or any equivalent arrangement.

LEVEL OF ORDINARY SKILL IN THE PERTINENT ART

The third Graham step is to resolve the level of ordinary skill in the pertinent art. The district court did not specify what he considered the pertinent art, nor the level of ordinary skill therein, although he made reference to “capabilities of persons working in the art to solve routine engineering problems.” The parties have not addressed argument to this point. We view the pertinent art as the design of trucks and trailers and the level of ordinary skill as that of an educated engineer with experience in such design, although not an expert in aerodynamics.

We are to consider whether the Saunders solution would have been obvious to a hypothetical person addressing the same problem, possessing that degree of skill, and having all the prior art at hand. See Republic Industries, Inc. v. Schlage Lock Co., 592 F.2d 963, 975 (7th Cir. 1979).

RELEVANT SECONDARY CONSIDERATIONS

Under Graham, secondary considerations such as commercial success or failure of others to solve the same problem may have relevancy.

Saunders’ device, manufactured by plaintiff Rudkin-Wiley, has enjoyed considerable commercial success. Sales increased to 4,000 units per year, 1969 to 1973. They almost doubled in 1974 and rose to 12,000 in 1976, the year before trial. Originally Saunders had tried unsuccessfully to interest six major truck manufacturers in his device. They all now frequently buy the *1208 device from Rudkin-Wiley for their customers.

The district court found that “any commercial success achieved by R-W [RudkinWiley] was the result of sales promotion and of outside factors, such as the OPEC oil crunch and government pressure on truck manufacturers.” 435 F.Supp. at 302. Mr. Wiley had testified to relatively small expenditures for promotion in the earlier years, but $35,000 in the years 1973-75 and $90,000 in 1976. He also indicated that sales “took off” at the time of the 1973 OPEC oil crunch. The court’s finding is not clearly erroneous.

Because of these circumstances the commercial success is less likely to raise the inference that many skilled people had addressed the problem over time without finding the solution.

The University of Maryland Wind Tunnel Operations Department conducted its tests of Trailmobile Trailers in 1953. The “tests were conducted to measure the drag of standard tractor trailer combinations and to explore methods of drag reduction.” Various modifications of tractor-trailers were tested to see whether and to what extent they would reduce wind resistance. One of the proposals for future action was to achieve a more streamlined body. No attempt was made to test a device similar to Saunders’ baffle. We are not shown the qualifications of those in charge of the tests, but it seems safe to assume they possessed a substantial degree of pertinent skill. The Saunders device was not obvious to them although they were addressing the problem for which we now know it was a significant solution.

Mr. Stamm appears to have been knowledgeable, and addressed the problem, but failed to produce the Saunders solution.

Plaintiff’s witness Smith, an aerodynamic expert and University professor, testified:

‘It’s not just another windshield. It is the neat combination of so deflecting the air, so positioning it, and so sizing it that it doés exactly this kind of thing [pushing the flow up so it comes over the top of the trailer] at its optimum position . .. So it is an interesting and subtle kind of trick here to have used a turbulent, high drag device and end up with a net savings.’

We note comments of the district court, 435 F.Supp. at 301-02 suggesting that he may have assumed that when an invention is a combination of old elements a synergistic effect must be identified before the invention can be found not to have been obvious. That is not an appropriate requirement. Republic Industries, Inc. v. Schlage Lock Co., 592 F.2d 963, 969 (7th Cir. 1979).

A patent is to be presumed valid, 35 U.S.C. § 282, but the presumption “does not exist against evidence of prior art not before the Patent Office.” Republic Industries, 592 F.2d at 973.

Rix, a Britain patent, was cited by the Examiner. The other bug deflector patents were not cited, but they do not add significantly to Rix. Huet was cited, and Lillie was no more pertinent than Huet. Hornke disclosed a slanting shield for luggage carrier to be located on the roof of an automobile. It has no bearing on the question whether Saunders’ invention was obvious.

This leaves the two demonstrations of scientific principles: (1) the da Vinci drawing indicating eddies of low pressure behind a baffle in a stream of water, and (2) the Paris experiment indicating that the spacing between a pair of bodies or the size of one body relative to the other will vary the drag on the second body. There is nothing to suggest that the Examiner had these principles in mind.

It is our judgment, however, that these two elements of prior art neither alone nor in conjunction with the other elements before us demonstrate that the Saunders devices would have been obvious to a person of ordinary skill in the pertinent art. The parties to this action have not argued that the considerations as to validity of the ’131 patent are significantly different from those involved in the ’876 patent. Although our analysis has been in terms of the ’876 patent for tractor-trailer combinations, we *1209 also conclude that the ’131 patent has not been shown to be invalid.

INFRINGEMENT

The district court, having concluded the patents were invalid, dealt only “briefly” with infringement. Argument by the parties on this issue was similarly brief.

We have reviewed the court’s findings and conclusions dealing with two aspects of the issue of infringement, finding error as to one, and no error as to the second. In a third, we conclude that the district court’s assertion that the doctrine of file wrapper estoppel “requires the conclusion that Defendants have not infringed Plaintiffs’ patents” is at least overboard.

Under the circumstances we want to make it clear that the district court’s opinion, and findings and conclusions as to infringement, other than specifically addressed and found not to be in error in this opinion, are not law of the case, governing further proceedings.

Claims 1 and 2 of ’876 included the recitation that the baffle is “forwardly con-vexed in a horizontal plane.” The district court found that the Air-Flo accused device is “forwardly concaved in a horizontal plane.” We do not agree.

The Air-Flo shield is similar in shape to a snow plow, with its point toward the front. Each side portion or shank is curved, so that each shank, taken separately, could be termed forwardly concave in a horizontal plane. Considered as a unit, however, the point is well forward of the two ends.

The only testimony addressed to the issue was that of plaintiffs’ witness Smith. He testified that “[g]rossly, it’s ‘forwardly eon-vexed. Locally, it is concave ... But grossly, it certainly is ‘forwardly convexed.’ ” The record includes pictures and drawings of the Air-Flo shield. We consider that the Air-Flo shield is “forwardly convexed in a horizontal plane” and the district court’s finding that the shield is forwardly con-caved is clearly erroneous. Thus this aspect of the Air-Flo shield does not save it from infringing claims 1 and 2 of ’876.

Claims 3 and 4 of ’876 included the recitation that the baffle “[have] its lower edge in substantially air impervious relation [with the cab roof].” The district court found that the accused device, spaced from four to five and one-half inches above the cab, does not have its lower edge in air impervious relation with the cab roof. We agree.

It is undisputed that the Air-Flo shield is mounted with its lower edge a few inches above the roof of the tractor. Defendants argue that because of the gap, their shield cannot be said to be in substantially air impervious relation with the roof. Plaintiffs argue that because little or no air actually flows front to rear through the gap when the tractor is moving forward, the structure is substantially air impervious.

A dictionary definition of impervious is “Not allowing entrance or passage through: impenetrable.” Webster’s New International Dictionary, 3d ed. at 1134. The ’876 specifications do not define the term nor otherwise explain or discuss it.

Plaintiffs’ witness Smith explained his understanding of the term “air impervious” as follows:

It means that the surface — the bottom is not truly — truly attached or absolutely impervious like exactly matched to the top of the cab of the truck.

But it means that as far as the air is concerned, there is little air flow, or maybe no air flow through there, that it just acts impervious to the air.

Smith described tests showing that a gap of up to six inches between the tractor roof and defendants’ device would not have any effect upon its eddy-creating and drag-reducing effects. Six inches was larger than the gap recommended by defendants for use of its product. When asked whether there would be any passage of air between the deflector and the roof when an Air-Flo device was mounted up to six inches above the tractor roof, Smith responded: “No. It wouldn’t significantly change the flow. . .. The air flow passing through is not sufficient enough to change the pattern ... *1210 The eddy and the flow up and over the top of the trailer.” Smith also testified that at certain levels of gap there was a flow of air from the rear to the front of the gap.

Given the ordinary meaning of “impervious,” the lack of any insight provided in the patent into the concept of “substantially air impervious relation,” and given the Smith testimony, we think the district court did not err in finding that the Air-Flo shield was not mounted in a substantially air impervious relation to the tractor roof. Accordingly, Air-Flo did not infringe claims 3 and 4 of ’876.

The judgment appealed from is reversed, and the cause remanded for further proceedings consistent with this opinion. To avoid any question whether the further proceedings will be a new trial for the purpose of Circuit Rule 18, we direct that the further proceedings be heard by Judge Allen Sharp, who conducted the original trial.

. Saunders v. Air-Flo Co., 435 F.Supp. 298 (N.D.Ind.1977).

. The district court also ruled against plaintiffs’ claim of trademark infringement. That ruling was not appealed.

. 435 F.Supp. at 299-303 (citing 35 U.S.C. § 102 (as to anticipation) and § 103 (as to obviousness)). The memorandum opinion was reported. Findings of fact and conclusions of law also filed were not reported.

. 435 F.Supp. at 304-06.

. The ’876 patent made the following claims:

(1) In combination with a tractor-trailer vehicle having a gap between the tractor and the trailer, said tractor being of less height than said trailer, an airflow deflecting baffle mounted to extend above the cab roof of said tractor for diverting the airflow relatively widely in a manner to avoid entry of the air stream into said gap thereby creating at least one low pressure air eddy in said gap to reduce the frontal air pressure against the trailer, said diverted air reattaching to the trailer at points spaced rearwardly of said gap, said baffle being inclined vertically rearwardly and forwardly convexed in a horizontal plane; positioned a distance from the front of the trailer equal to approximately 0.7 the half width of the trailer end of a height substantially 0.7 of the difference in height between the tractor cab roof and the roof of the trailer.

(2) In combination with a tractor-trailer vehicle having a gap between the tractor and the trailer, an air flow deflecting shield comprising a baffle mounted to extend above the tractor cab roof, said baffle being vertically inclined rearwardly and forwardly convexed in a horizontal plane, said baffle having a predetermined height substantially 0.7 of the difference in height between said cab roof and the trailer roof, said baffle being positioned a distance from the front of the trailer equal to approximately 0.7 the half width of the trailer.

(3) In combination with a tractor-trailer vehicle having a gap between the tractor and the trailer, said tractor having a cab with the roof thereof of less height than said trailer, a substantially solid upstanding air flow deflecting baffle mounted to extend above the cab roof and having its lower edge in substantially air impervious relation therewith, for diverting the substantially entire air flow relátively widely in a manner to avoid entry of the air stream into the gap, thereby creating at least one low pressure air eddy in said gap to reduce the frontal air pressure against the trailer, the diverted air reattaching to the trailer at points spaced rearwardly of said gap, said baffle being of a height between 0.5 and 0.9 of the difference in height between the tractor cab roof and the roof of the trailer, and being positioned a distance from the front of the trailer equal to between 0.3 and 2.0 times the half width of the trailer.

(4) The structure of claim 3 wherein said baffle is positioned immediately adjacent the rear of the cab roof.

. The T31 patent made the following claims:

(1) A means of reducing linear wind resistance in combination with a single chassis vehicle including a cab having a windshield, an engine compartment having its top terminating below said windshield and a body having a height in excess of the height of said cab and having its front edge closely juxtaposed to the rear of said cab leaving no appreciable air gap therebetween, said means comprising an upstanding baffle mounted adjacent the forward edge of the cab at the top of the windshield so dimensioned as to deflect the airstream impacting the windshield and baffle over the top of said front edge of the body, the airstream reattaching to the top *1203 of the body at a point rearwardly of said forward edge in a smooth linear flow defining a low pressure air eddy on the cab roof between said baffle and said front edge of said body, the height of the windshield plus the height of the baffle being equal to between 0.6 and 0.8 of the height of said body above said engine compartment.

(2) The structure of claim 1 wherein said baffle extends transversely of the cab.

(3) The structure of claim 1 wherein said baffle is inclined rearwardly vertically.

(4) The structure of claim 1 wherein the ends of said baffle are inclined rearwardly relative to the transverse plane of the baffle.

(5) The structure of claim 4 wherein said inclined ends are arcuate and forwardly con-vexed.

(6) The structure of claim 1 wherein, when h equals the combined height of said windshield and said windshield and said upstanding baffle and H equals the height of said body above a line extending along an extension of the line of division of air between the windshield and the said engine compartment, h equals 0.7H.

(7) The structure of claim 1 wherein the half-width of said baffle is substantially equal to 0.7 of the half-width of the body.

(8) The structure of claim 1 wherein the distance of said baffle from the front of said body is substantially equal to 0.7 of the half-width of the body.

(9) The structure of claim 6 wherein the half-width of said baffle is substantially equal to.0.7 of the half-width of the body.

(10) The structure of claim 6 wherein the distance of said baffle from the front of said body is substantially equal to 0.7 of the half-width of the body.

(11) The structure of claim 9 wherein the distance of said baffle from the front of said body is substantially equal to 0.7 of the half-width of the body.

. A fairing is a smooth contoured surface designed to provide streamlining by keeping air flow attached to its surface from the tractor roof to the trailer roof.

. Appellee also views these terms and “baffle” as interchangeable because all of these devices “serve the function of deflecting and regulating air flow .... ” Appellee’s brief at 13.