OPINION OF THE COURT
MANSMANN, Circuit Judge.In this diversity matter we are asked to examine the substantive tort law of the Commonwealth of Pennsylvania in order to determine the role of expert testimony and the so-called “fiber drift” theory in proving causation in cases of alleged asbestos-related injury. Because the Supreme Court of Pennsylvania has not addressed the question before us, our task is to predict the position which that court would take in resolving this dispute. We conclude that the fiber drift theory, when grounded in the standards set forth in Pennsylvania decisions, is a factor to be considered in determining whether a plaintiff has adduced evidence of causation sufficient to withstand a motion for summary judgment.
I.
These actions, brought by eight plaintiffs and their spouses, were part of a series of complaints filed in the U.S. District Court for the Eastern District of Pennsylvania on behalf of approximately 600 former employees of the former Firestone Tire and Rubber Company Plant in Pottstown, Pennsylvania (“Firestone”). These suits, a product of the National Tire Workers Litigation Project, named as defendants thirty manufacturers and suppliers of asbestos products to which the plaintiffs were allegedly exposed while employed at Firestone. The plaintiffs claim personal injury as a result of exposure to various asbestos-containing products including insulation, gaskets, friction materials, fillers, talc and soapstone. The plaintiffs allege that their asbestos-related personal injuries were caused in substantial part by exposure to asbestos shed by brake linings manufactured by Bendix, the predecessor to defendant Allied-Signal, Inc., cement manufactured by Eagle-Picher Industries, Inc., talc containing asbestos contaminants manufactured by Vermont Talc, Inc. and gaskets and packing manufactured by John Crane, Inc.
The cases of these eight plaintiffs were scheduled for two separate trials, one to begin on September 25, 1989, and the other on October 2, 1989.1 Between September 13, 1989 and October 4, 1989, four defendants, Allied Signal Inc., Eagle-Picher Industries, Inc., John Crane, Inc. and Vermont Talc, Inc. filed motions for summary judgment. The motions of Defendants Eagle-Picher and John Crane were filed in the cases of Plaintiffs Reimert, Yourkavitch, Moyer and Davis only. All of the motions focused on the issue of causation. Specifically, the defendants contended that despite extensive discovery, the plaintiffs had failed to establish the requisite nexus between their alleged injuries and particular asbestos-containing products manufactured by the defendants.
The plaintiffs responded to the motions for summary judgment, maintaining that their own deposition testimony, the testimony of product identification witnesses, and the testimony of expert scientific witnesses would provide sufficient evidence of use and deterioration of the defendants’ asbestos-containing products throughout the Firestone plant for a jury reasonably to *365infer that fibers from these products contributed substantially to the plaintiffs’ injuries.
In arguing that the causation issue should be submitted to the jury, the plaintiffs placed principal reliance upon the testimony of scientific witnesses. These witnesses testified that asbestos-containing products of the type manufactured by the defendants would have been capable of emitting asbestos fibers in the Firestone plant. These fibers could have been carried by air currents and other means to various parts of the plant where they may have been inhaled by the plaintiffs, thus contributing to the alleged asbestos-related injuries. This theory that airborne asbestos fibers contribute to bystander disease is characterized as the “ambient air” theory, “re-entrainment,” or as “fiber drift.” For the purposes of this opinion, we adopt the last term.
On November 1, 1989, the district court granted the defendants’ summary judgment motions with respect to all of the plaintiffs,2 stating Pennsylvania substantive law as follows:
To withstand summary judgment, plaintiffs must present some competent evidence that the defendant’s product was present in the workplace, that the plaintiffs worked in the vicinity of the product’s use on a regular basis and inhaled asbestos fibers shed by defendant’s products.
In re Tire Worker Asbestos Litigation, Nos. 88-4702 and 88-4703, slip op. at 5, 1989 WL 133650 (E.D.Pa. Nov. 1, 1989), citing Eckenrod v. GAF Corp., 375 Pa.Super. 187, 544 A.2d 50 (1988). The court found that “none of the eight plaintiffs has produced any evidence whatsoever that he was exposed to any Eagle-Picher, Bendix/Allied Signal, Vermont Talc or John Crane asbestos-containing product at the Firestone plant.” Slip op. at 5-6. The totality of the evidence submitted was summarized as follows: (1) the plaintiffs could not personally identify any product to which they were allegedly exposed; (2) “[a]t best their product identification witnesses might have seen some of the plaintiff-husbands using unidentified thermal insulation, brake and talc products;” slip op. at 7; and (3) the asbestos fiber drift theory advanced by the plaintiffs’ experts was “speculative” since it could establish no more than that the “plaintiffs may have been exposed to the mixing and conglomeration of the airborne asbestos fibers, some of which may have been shed by defendants’ asbestos-containing products.” The district court concluded that the evidence presented was insufficient to establish causation under the prevailing law of Pennsylvania. Slip op. at 8. The order granting summary judgment was certified for appeal pursuant to 28 U.S.C. § 1292(b).3
On November 10, 1989, the plaintiffs filed a motion to reconsider and to vacate the order granting summary judgment. In this motion the plaintiffs contended that the court had issued its November 1, 1989 memorandum opinion and order “without having the benefit of considering plaintiffs’ scientific experts’ deposition testimony.” The deposition testimony of the experts and an additional affidavit were appended to the motion. On November 30, 1989, prior to the district court’s ruling on the motion to reconsider, the plaintiffs moved for permission to appeal. We declined to *366accept certification of the Order under 28 U.S.C. § 1292(b).
On December 4, 1989, the district court denied the plaintiffs’ motion to reconsider, ruling that the expert evidence presented by the plaintiffs was not sufficient under Pennsylvania law to defeat a summary judgment motion; the plaintiffs had not offered adequate evidence of product identification:
The court remains fully convinced that the plaintiffs can establish the asbestos fiber drift theory through the testimony of experts. However, these experts, no matter how many they might number and no matter what their testimony might be cannot convince the court ... that the asbestos fiber drift theory satisfies the substantive law of Pennsylvania.
In re Tire Worker Asbestos Litigation, Nos. 88-4702 and 88-4703, slip op. at 4-5 (E.D.Pa. Dec. 4, 1989). On December 6, 1989, summary judgment was again granted to the defendants and the order was certified as final pursuant to Fed.R.Civ.P. 54(b). In re Tire Worker Asbestos Litigation, Nos. 88-4702 and 88-4703, slip op., 1989 WL 150597 (E.D.Pa. Dec. 6, 1989).
II.
We note at the outset that the number of asbestos cases which have been filed in this jurisdiction and across the country is formidable. “More than 30,000 asbestos personal injury claims [had been] filed nationwide by 1986 and an additional 180,000 are projected to appear on court dockets by the year 2010.” In re Asbestos Litigation, 829 F.2d 1233, 1235 (3d Cir.1987), cert. denied, sub nom., Owens Illinois Inc. v. Danfield, 485 U.S. 1029, 108 S.Ct. 1586, 99 L.Ed.2d 901 (1988). As the district court pointed out, some 1,000 of the cases are pending in the Eastern District of Pennsylvania alone. There is no body of federal law applicable to these claims; they are governed by the substantive tort law of the various states and are litigated under the theories of negligence, warranty and strict liability. Recognizing that our assessment of the fiber drift theory and its place in the law of Pennsylvania will have an impact beyond the confines of this appeal, we turn our attention to the issue presented.
This appeal requires that we determine the •prima facie case that a plaintiff, allegedly injured as a result of exposure to asbestos fibers, must establish with regard to a defendant’s product in order to survive a motion for summary judgment. We focus particularly upon the fiber drift theory advanced by the plaintiffs’ expert scientific witnesses in order to determine whether that theory may be invoked to establish causation under the products liability law of the Commonwealth of Pennsylvania.
Our review of a district court’s grant of summary judgment is plenary. We apply the same test for summary judgment as that applied by the district court “with all evidentiary inferences resolved in a light most favorable to the non-moving party.” Aronow Roofing Co. v. Gilbane Bldg. Co., 902 F.2d 1127, 1128 (3d Cir.1990). Summary judgment is properly granted where the non-moving party “fails to establish the existence” of an element essential to the case. Celotex Corp. v. Catrett, 477 U.S. 317, 322, 106 S.Ct. 2548, 2552, 91 L.Ed.2d 265 (1986). We review the district court’s refusal to reconsider its entry of summary judgment for abuse of discretion. Harsco Corp. v. Zlotnicki, 779 F.2d 906, 909 (3d Cir.1985), cert. denied, 476 U.S. 1171, 106 S.Ct. 2895, 90 L.Ed.2d 982 (1986).
In order to assess the propriety of the district court’s grant of summary judgment in this matter, we turn first to the substantive tort law of Pennsylvania.
III.
Under Pennsylvania law, the plaintiff in a products liability action bears the burden of demonstrating that a specific defendant is responsible for the harm alleged. Berkebile v. Brantley Helicopter Corp., 462 Pa. 83, 337 A.2d 893 (1975). Proof of causation involves two elements: proof of cause in fact and proximate cause. Cause in fact or “but for” causation requires proof that the harmful result would not have come about but for the conduct of the defendant. Proximate cause, in addi*367tion, requires proof that the defendant’s conduct was a substantial contributing factor in bringing about the harm alleged. Where the relevant facts show either that the defendant was not responsible for the injury, or that the causal connection between the defendant’s negligence and the plaintiffs injury is remote, the question of causation is decided by the court as a matter of law. See Conti v. Ford Motor Co., 743 F.2d 195, 198 (3d Cir.1984), cert. denied, 470 U.S. 1028, 105 S.Ct. 1396, 84 L.Ed.2d 784 (1985) and Morena v. South Hills Health System, 501 Pa. 634, 462 A.2d 680 (1983).
What must be shown in order to establish the requisite causal connection between a plaintiff’s asbestos-related injury and the defendant’s asbestos-containing product is central to resolution of this case. In finding that the plaintiffs here have failed to establish a nexus between their alleged injuries and any of the defendants’ products, the district court looked primarily to the Pennsylvania Superior Court’s decision in Eckenrod v. GAF Corporation, 375 Pa.Super. 187, 544 A.2d 50, alloc. denied, 520 Pa. 605, 553 A.2d 963 (1988).
In Eckenrod the plaintiff’s decedent had been employed as a millwright and maintenance pipefitter in the Wallace Run facility of Babcock and Wilcox. After more than 20 years of service, he died of lung cancer. The plaintiff alleged that her husband’s death resulted from exposure to asbestos-containing products manufactured by several defendants. Summary judgment motions made by a number of defendants were granted based on the lack of product identification.
While co-worker affidavits showed that Eckenrod had been exposed to asbestos products, none of the witnesses clarified the proximity of the products to the deceased or were able to establish that the defendants had manufactured or supplied the products used. The only specific product identification testimony was contained in depositions of asbestos products distributors and one plant storeroom employee. While this testimony established that the defendants’ asbestos products had been sold to Babcock and Wilcox, it did not show that the specific product had been used or that the decedent had ever worked in proximity to the product.
In affirming the grant of the defendants’ motion for summary judgment, the Superi- or Court evaluated the evidence presented:
Whether direct or circumstantial evidence is relied upon, our inquiry, under a motion for summary judgment, must be whether plaintiff has pointed to sufficient material facts in the record to indicate that there is a genuine issue of material fact as to the causation of decedents’ disease by .the product of each particular defendant. Schmidt v. Johns Manville Corp., No. 80-3339 slip op. (D.Md. Nov. 30, 1982). Whether a plaintiff could successfully get to the jury or defeat a motion for summary judgment by showing circumstantial evidence depends upon the frequency of the use of the product and the regularity of plaintiff’s employment in proximity thereto. Id. ... We acknowledge that the facts establish that the decedent on occasion was exposed to asbestos; there is no evidence, however, as to the regularity or nature of the decedent’s contact with asbestos. Moreover, there is no testimony establishing that [the decedent] worked with asbestos manufactured or supplied by [the defendants]. The mere fact that [defendants’] asbestos products came into the facility does not show that the decedents ever breathed those specific asbestos products or that he worked where these asbestos products were delivered. ...
544 A.2d at 53.
What has come to be known in Pennsylvania as the Eckenrod “frequency, regularity and proximity test” was developed in the U.S. District Courts in Maryland and adopted by the Court of Appeals for the Fourth Circuit in Lohrmann v. Pittsburgh Corning Corp., 782 F.2d 1156, 1162 (4th Cir.1986). In Lohrmann, a plaintiff ship-worker argued that since he had presented evidence placing a company’s asbestos-containing product in the workplace during the time of his employment, it was for the jury *368to determine whether that product contributed as a proximate cause to his disease. The court concluded that:
Such a rule would be contrary to the Maryland law of substantial causation. In dealing with a workplace as large as a shipyard, certain of the district judges have adopted a ... standard for evaluating the sufficiency of the evidence of exposure....
To support a reasonable inference of substantial causation from circumstantial evidence, there must be evidence of exposure to a specific product on a regular basis over some extended period of time in proximity to where the plaintiff actually worked.
Id. at 1162-63. The court, affirming the entry of a directed verdict for a number of the defendants, explained that “there must be evidence of a reasonable and rational nature upon which a jury can make a rational inference that there is a causal connection between a defendant’s action and a plaintiff's injury.” Id. at 1163.
The Eckenrod frequency, regularity and proximity test of causation in asbestos litigation articulates the evidentiary requirements applied in earlier decisions interpreting Pennsylvania law, see Anastasi v. Pacor, Inc., No. 6251 (C.P.Phila.Co., March 8, 1983), aff'd, 349 Pa.Super. 610, 503 A.2d 44 (1985), and has been confirmed by another panel of the Superior Court in Samarin v. GAF Corp., 391 Pa.Super. 340, 571 A.2d 398 (1989), alloc. denied, 524 Pa. 629, 574 A.2d 71 (1990). In Samarin, the plaintiffs alleged injury as a result of exposure to asbestos dust released from products used to control temperature in a Babcock and Wilcox plant facility. After reviewing the plaintiffs’ testimony, interrogatory answers, and affidavits, the court upheld an award of summary judgment in favor of the defendants on the ground that the plaintiffs had failed to establish a sufficient connection between their alleged injuries and the defendants’ products.
The Eckenrod standard has been cited by federal decisions interpreting Pennsylvania law4 and a similar nexus requirement has been upheld in other jurisdictions.5
*369IV.
It was against the standards developed in Eckenrod and Samarin that the district court evaluated the evidence bearing upon causation in this case. Much of the evidence offered centered on the physical characteristics of the Firestone plant.
The Firestone tire-manufacturing facility, in which all of the plaintiffs were employed, operated from the 1940’s until 1980 when the plant was sold. During the relevant period, the facility consisted of at least six buildings spread over more than 200 acres. The actual manufacture of tires took place in what is referred to as the “tire building.”
The tire building consisted of three levels covering a minimum of 862,000 square feet of floor space. The building’s three levels were designated by their height in feet above sea level, the 155’ level, the 167’ level, and the 180’ level. The 155’ level housed the mill room, an open space with no partitions and covered 196,000 feet. The 167’ level contained the tube department and covered 258,000 square feet and the 180’ level covered 408,000 square feet. With the exception of a closed plastics area, there were no partitions, between departments on the 180’ level. The plaintiffs’ primary allegations of asbestos exposure related to their employment in this building.
At the time of the defendants’ summary judgment motions, the district court had before it three categories of evidence offered to establish the plaintiffs’ exposure to the defendants’ products: (1) the plaintiffs’ own direct testimony; (2) the testimony of product identification witnesses; and (3) expert opinion. Because the probative value of each of the categories is fact-specific, we will summarize the substance of each category of evidence seriatim.
A. The Plaintiffs’ Direct Testimony
1.John Davis — Mr. Davis worked at Firestone from 1968 to 1980. He worked in the mill room on the 155’ level for three years; this room was very dusty. He also worked for five years in the compound room on the 180’ level which was also dusty. Forklifts frequently drove into low-hanging pipes, causing ceiling material to fall on him. This had also happened in the mill room. Mr. Davis spent one year in the curing area on the 180’ level where there was a great deal of piping.
2. Ammon Moyer — Mr. Moyer was employed at the Firestone plant from 1946 to 1980. He worked for six months in the tread-ply assembly section on the 180’ level where air quality was “poor.” He spent three years as a machine operator on the 180’ level and was exposed to asbestos on steam pipes. Fork trucks struck the pipes at least once a week and the pipes were maintained every day, resulting in the release of light flaky dust. For 22 years, Mr. Moyer worked as a final inspector in an area adjacent to the curing process on the 180’ level. Pipes in the curing area were covered with asbestos and the air in the final inspection area was “dusty” from the asbestos and talc.
3. Kenneth Reimert — Mr. Reimert began working in the Firestone tire plant in 1960 as a serviceman in the mill room on the 155’ level. In 1964 he moved to the tread tubing area on the 180’ level. He worked on the same level from 1970 to 1980 as a tire builder. He described four years of working with soapstone which was sprayed onto the rubber above him and stated that there were asbestos pipes in all areas in which he worked. The pipes were being maintained “constantly” and Mr. Rei-mert sometimes performed emergency pipe-wrapping jobs. Mechanics made repairs to the building machines “right in front of him.”
4. Stanley Yourkavitch — Mr. Yourka-vitch began work at Firestone in 1946 in the curing department on the 180’ level. From 1962 to 1980, he worked in final inspection on the same level. He testified to using soapstone during his employment *370and to having worked in the vicinity of asbestos-covered pipe for at least 16 years.
5. Edwin Beidler — Mr. Beidler was employed at the Firestone plant from 1960 to 1980 as a pipefitter; from 1961 to 1980, he worked primarily in the chemical plant. He spent one year exclusively in the tire plant although he was later called upon to make repairs there on a regular basis. As a pipefitter Beidler often removed asbestos cement and insulation. The tube room, where he sometimes worked, was always dusty because of talc use. He also worked with asbestos sheeting, gaskets, and packing and saw repairs made to the brake linings of wheeled vehicles.
6. Willard Klein — For 34 years Mr. Klein worked in the tire curing and building area on the 180’ level. He claimed exposure to asbestos from overhead steam pipes.
7. James Sgro — Mr. Sgro began working at Firestone in 1970 and spent seven years in the tube room on the 167’ level where he was exposed to soapstone. From 1977 to 1980 he worked in the tire building area on the 180’ level. He was exposed to covered steam pipes throughout his 10 year employment.
8. Larry Stopfel — Mr. Stopfel spent seven years in the stock-cutting area on the 180’ level where he was exposed to soapstone and another seven years running a mixer and operating a forklift. Asbestos insulation was used in all of the areas in which Stopfel worked.
B. Co-Worker Testimony Regarding Product Identification
1. The Products of Allied-Signal, Inc. (successor to Bendix) (“Allied/Bendix”)
Frank K. Elliott, Jr., a mechanic, testified in his deposition that pneumatic brakes with asbestos-containing linings were used on truck-tire building machines during the period from 1946 to 1976. Brake housings were impressed with the word “Bendix.” “Bendix” shipping slips accompanied replacement assemblies. Overhead cranes in two departments used Bendix brakes. Elliott performed brake repairs on bias machines in Plaintiff Moyer’s presence at least 25 times between 1975 and 1980 but Elliott was unable to identify the manufacturer of the brake linings. Elliott made one repair of a Bendix brake in the presence of Plaintiff Yourkavitch.
Edward J. Brunke, a maintenance foreman, testified that Bendix brakes were used on electric motors on the calendars from 1957 to 1959.
Frank K. Leininger, Jr., a mechanic, testified that Bendix brake linings were used on Spadone cutters from 1967 to 1980 and that the brake linings contained asbestos.
Paul A. Guthrie, a mechanic, testified that the name “Bendix” was impressed on the brake housing of forklift trucks.
Allied’s response to the evidence presented — Firestone’s records reveal no purchases from Bendix or its successor, Allied-Signal. Even if there were a small number of Bendix brake linings used at Firestone, records showed that substantially all of the brake linings used in the plant were manufactured by Raybestos-Manhat-tan. This was uncontroverted. Allied contended that the testimony offered failed to connect Bendix products to any of the eight plaintiffs with the exception of one instance in which a new Bendix brake was installed in Mr. Moyer's presence. Allied argues that the testimony did not show that a Bendix brake had been removed from the machine under repair. Allied also stated that “Bendix” on the brake housings did not establish that Bendix had been the manufacturer of the brake linings.
2. The Products of Eagle-Picher Industries, Inc.
John Helmbold, a pipe coverer, testified that from 1964 to 1980 he was regularly assigned to the tire building. He testified that he had used Eagle-Picher “66” asbestos cement to insulate pipes in the curing areas, the cafeteria, and on piping leading to the plastic calendar. Helmbold was unable to pinpoint exactly when he used the Eagle-Picher cement in the curing area, how much he used, or who was present when he used it. He testified that no one *371was present on other occasions when he used the cement. He used other manufacturers’ cements more often than he used the Eagle-Picher product.
Sterling Haydt testified that he had seen Eagle-Picher cement in the plant between 1960 and 1980. He never used it himself and did not identify any of the plaintiffs as having used the cement or as having been present during its use.
Ralph Minotto, a pipefitter, recalled that One-Cote Cement, an Eagle-Picher product, was used to finish joints in the late 1960’s. Mr. Minotto did not testify that he used the cement and did not connect the product’s use with any of the plaintiffs.
Anthony Panfile, another pipefitter, testified that Hylo pipe covering, another Eagle-Picher product, was used between 1947 and 1960. Mr. Panfile could not recall where this covering had been used and none of his testimony connected this product with the plaintiffs.
Nicholas S. Saccomanno, the plants’ sole pipe coverer in 1963 and 1964, was permanently assigned to the chemical plant in 1964. He assisted in the tire building approximately six times a year for six years. Saccomanno stated in his deposition that Eagle-Picher “66” cement and other cements were in the tire and chemical plants from 1966 through 1970. Saccomanno testified to the specific use of Eagle-Picher cement only on the fittings to tire curing stands in the pot heater basement over 3 to 4 months in the 1963-1964 period.
Eagle-Picher’s Response — The product identification testimony established no more than the mere presence of unquantifiable amounts of Eagle-Picher products. The testimony did not show where any Eagle-Picher product was installed in relation to any plaintiff, how often Eagle-Picher products were used in relation to the numerous other thermal insulation products used at Firestone, whether the products contained asbestos, when any Eagle-Picher thermal insulation products may have been damaged, repaired or removed, or whether any plaintiff was exposed to any dust from these products over a significant period of time.
3. The Products of John Crane, Inc.
Pipefitter Anthony Panfile testified that between 1948 and sometime in the 1960’s, he used “Crane” gaskets and packing on high pressure all-temperature steamlines, vessels, pumps and valves on a “plant-wide” basis. Fibers would come off “a lot of the gaskets” when scratched off the flanges. Panfile knew that Plaintiff Bei-dler repacked valves with “Crane” packing.
Edward J. Brunke, a maintenance foreman at the tire plant, recalled seeing “Crane” asbestos packing and gaskets from 1957 to 1959 and 1961 to 1964. Pipe-fitter Harry L. Morgan identified John Crane gaskets and packing, particularly sheet packing, as having been used in the curing area in 1964. Plaintiff Beidler was present. These products were also used in the curing and building areas in 1969 and 1970 and again in the middle 1970’s and were used in performing central maintenance in the late 1970’s. The sheet packing was used with Plaintiff Beidler in 1969.
Pipefitter Carl Janiszewski testified that he had used John Crane spool form graphite impregnated asbestos packing from 1969 to 1980 on a variety of pumps and valves in the tire and chemical plants. A box of John Crane packing indicated that it contained asbestos. He was unable to connect John Crane packing to any of the plaintiffs.
Frank Fomado, a maintenance supervisor, testified that John Crane asbestos sheet packing was used to make gaskets in the chemical plant and that he had seen Plaintiff Beidler working with gaskets at that location.
John Crane, Inc.’s Response6 — In addressing the testimony of the product identification witnesses, John Crane, Inc. notes that it is the corporate successor to Crane Packing Company and John Crane-Houd-aille, Inc. These companies did sell some *372packing and gasket materials containing encapsulated asbestos which did not emit respirable fibers. John Crane, Inc. further notes that its products never contained only the identifying word “Crane” as “Crane” was the name of a competing corporation which also sold gaskets and packing materials containing asbestos.
No plaintiff, other than Beidler, offered evidence relating to exposure to gaskets or packing or mentioned observing dust or airborne debris emanating from gasket or packing products. The purchasing agent for the Firestone plant testified that Firestone purchased no asbestos-containing materials manufactured by the Crane Packing Co. (the name of John Crane, Inc.’s corporate predecessor during the time of Firestone’s Pottstown operations). Gasket materials were purchased from several other companies.
The testimony of Mr. Janiszewski and Mr. Morgan with regard to John Crane packing did not establish that asbestos-containing packings resulted in emission of airborne dust or debris. Morgan did not know whether the packings which he used contained asbestos. Witnesses Panfile, Brunke and Formado testified to seeing “Crane” gaskets. Panfile and Brunke did not know whether these “Crane” materials contained asbestos and Formado conceded that he had been aware that “Crane” and John Crane, Inc. were separate manufacturers of packings and could not say whose products he had seen at Firestone. Furthermore, none of the plaintiffs involved in John Crane Inc.’s summary judgment motion worked in the chemical plant where Formado had seen “Crane” products.
No product identification witness placed any of the plaintiffs, other than Beidler, in the vicinity of the gaskets.
4. The Products of Vermont Talc, Inc.
H. M. Royal, a distributor for Vermont Talc, showed sales of talc to Firestone between 1963 and 1980.
The plaintiffs’ expert mineralogist, Arthur Rohl, supplied a report and affidavit that Vermont Talc’s talc product contained 0.2 to 0.5 per cent asbestiform minerals, which would have become airborne and res-pirable when the talc was used.
Ralph Minotto, Frank Elliott and William Pfingsten testified to the extensive use of talc and soapstone in the tube mill rooms.7
Vermont Talc’s Response — Duncan Ogden, Vermont Talc’s expert witness, testified that in 1972 Vermont Talc instituted a program to test for the presence of asbestos in its talc products. From 1972 forward, no asbestos was found. Geologic deposits mined prior to 1972 were identical to those mined after 1972 and were, therefore, presumed not to contain asbestos.
Vermont Talc draws a distinction between talc and soapstone, stating that they are, in reality, different products. Plaintiffs Reimert, Yourkavitch, Sgro, Stopfel and Davis testified regarding exposure to soapstone rather than talc. Plaintiff Moyer claimed exposure to talc or soapstone and Plaintiff Beidler claimed exposure to both. Plaintiff Klein did not allege exposure to either product. No plaintiff was able to identify Vermont Talc products.
C. The Plaintiffs’ Expert Testimony and the Fiber Drift Theory
In addition to the plaintiffs’ testimony and the testimony of product identification witnesses, the plaintiffs offered the testimony of several expert witnesses in order to establish the plaintiffs’ exposure to the defendants’ products. This testimony, in the form of reports and affidavits, was offered to prove that the defendants’ asbestos products regularly emitted asbestos fibers in the open and poorly ventilated Firestone plant. These fibers were said to have been carried by air currents and inhaled by the plaintiffs. We set forth here a summary of the “fiber drift” testimony of each expert witness.
*373Kenneth S. Cohen, Ph.D. — In an affidavit and a report, Dr. Cohen identified himself as a professional engineer and certified industrial hygienist who had participated in a survey of the Firestone plant on August 29, 1989. At the plant he observed “abundant evidence of extensive asbestos (and asbestos product) use [and] a variety of installation and repair processes which by their very nature release asbestos fibers. The vast network of asbestos insulated pipe systems, asbestos packed valves, asbestos sealed flanges, unions and other asbestos containing materials seen during our brief survey offers proof of the substantial nature of workers’ exposures.” He observed a “steam hammer” effect or vigorous shaking in the extensive piping which causes deterioration of insulating asbestos and a “salt shaker” phenomenon. The open space and cross-ventilation created a homogenizing effect throughout the plant, dispersing asbestos fibers from their source to bystanders. Dust samples showed evidence of talc and asbestiform fibrous structures. Nothing in this report tied the plaintiffs’ exposure to defendants’ specific products.
William J. Nicholson, Ph.D. — Dr. Nicholson, a Professor of Community Medicine at Mount Sinai Medical Center in New York, stated that he had supervised an industrial hygiene study of insulation work. In his report he noted that the application, repair, removal of or damage to asbestos containing products (insulation in particular) can release substantial quantities of respirable asbestos fibers which can travel on air currents throughout a workplace. He cited a study finding wide-spread dissemination of fibers during removal of fireproofing material from an aircraft carrier and stated that he, himself, had measured asbestos dissemination at a construction site where fibers were carried one-half mile. He cited a second study showing increased cancer risk for those working in the vicinity of, rather than directly with, asbestos-containing products. He was unable to calculate the concentrations of asbestos fibers which might have been present in the Pottstown plant as a result of dissemination from insulation materials.
Commenting on Dr. Cohen’s report, Dr. Nicholson concluded that evidence of steam hammer vibration in many of the pipes and evidence of repairs to insulation suggested that the tire plant contained many widely dispersed sources of asbestos which would provide substantial and ongoing exposure to workers throughout the plant. Exposure could also have resulted from repeated disturbance of asbestos-containing debris which had fallen to the floor and from recirculation of fibers through the ventilation system. Discussing talc, Dr. Nicholson concluded that “to the extent that talcs were contaminated, fibers from this source would have added to exposure from insulation materials.” He did not mention possible exposure from gaskets or brake linings.
It was on the basis of this record, summarized above, that the district court entered summary judgment in favor of the four defendants.
V.
On November 10, 1989, the plaintiffs filed a motion to reconsider the order granting summary judgment. The plaintiffs asserted that throughout the district court’s consideration of the motions for summary judgment, the plaintiffs had urged that decisions be deferred pending the conclusion of the plaintiffs’ experts’ depositions and had operated on the understanding that they would be given the opportunity to have these additional materials considered. The summary judgment motions were granted, however, prior to the court’s having had the opportunity to review the scientific experts’ deposition testimony. In the plaintiffs’ view, this testimony was necessary in order “to fully weigh the probative value of ‘asbestos fiber drift’ vis-á-vis the identification of asbestos-containing products within the traditional standards set forth in Eckenrod.”
The plaintiffs thus appended the deposition testimony of George B. Stanton, Jr., an industrial hygienist, Dr. William Nicholson, the professor of community medicine whose report was detailed above, and Dr. *374Alf Fishbein, a pulmonologist, to the motion for reconsideration. Affidavits executed by Stanton and Dr. Robin Kaplan, a physician, Board Certified in internal and pulmonary medicine, were also appended. In addition to the expert testimony, the plaintiffs offered the deposition testimony of William Pfingsten, an additional product identification witness.
In denying the plaintiffs’ motion for reconsideration, the district court did not refuse to review the testimony because it was presented in an untimely manner; the court concluded instead that the additional evidence submitted was not relevant under Pennsylvania law. For reasons which we will set forth in detail below, we conclude that it was an error for the district court to reject the proffered expert testimony insofar as it related to fiber drift.
In our view, Pennsylvania law requires that a district court exclude expert scientific testimony on the issue of fiber drift only where the other testimony presented does not show that the defendants’ products were used frequently in an area in which a plaintiff was regularly employed. Because we find that this factual foundation was established as to some products, we conclude that the district court should have examined the fiber drift evidence. We will summarize the excluded evidence in order to clarify our discussion of the fiber drift theory and causation under the law of Pennsylvania.
The deposition of William Nicholson— Relying on Dr. Cohen’s conclusions, excerpts of the plaintiffs’ depositions, portions of the deposition testimony of product identification witnesses and the recollections of former Firestone employees, Nicholson offered testimony regarding air movement in the Firestone facility. In his opinion, asbestos fibers from various sources would have circulated within the plant with “opportunity for widespread dissemination.”
The deposition and affidavit of George B. Stanton, Jr.—Stanton, a Professional Engineer, Certified Safety Professional and Certified Industrial Hygienist, reviewed the reports of Drs. Cohen and Nicholson in formulating his conclusions regarding the plaintiffs’ asbestos exposure at Firestone. He also took into account information provided by former engineers and workers at the Firestone plant, studied photographs, floor plans and videotapes of the facility, and reviewed the depositions of the plaintiffs and the product identification witnesses. Relying upon these sources of information, Stanton concluded that each of the plaintiffs had been exposed to asbestos fibers released by products manufactured by John Crane, Inc. and Eagle-Picher Industries, Inc., that “plaintiffs regularly employed on the 180’ level of the tire plant between 1948 and the late 1970’s were frequently exposed to asbestos fibers emitted by Bendix products, and that all plaintiffs who regularly worked on the 155’ and 167’ levels between 1951 and 1978 were frequently exposed to asbestos contaminants in Vermont Talc, Inc.’s products.”
Stanton concluded that much of the exposure resulted from fiber drift. He described fiber drift as follows:
[I] say the exposure was continuing because whatever dust settled to the floor or settled to equipment surfaces or whatever, then was kicked up once again by people brushing against the equipment surface or walking or the industrial trucks that were used in this plant or whatever motivate [sic] or redisbursed [sic] the settled asbestos dust into the air.
... Beyond that, all of these forces created attrition of the asbestos fibers making the larger ones smaller ones.
Stanton testified that given air flow conditions in the Firestone plant, contaminated air was recirculated and recirculation, combined with materials handling, could cause a person 150 feet from a source of asbestos fiber to be exposed to the same extent as another person standing ten feet from the source.
Stanton supplied no scientific data to support this statement and was unable to say what percentage of fibers was recirculated or to quantify the percentage of fibers removed from the air by the ventilation systems.
*375Stanton was not able to identify studies validating the fiber drift theory and had conducted no studies of his own. He concluded that “[i]t had to be that way.”
Stanton admitted that he did not rely upon measurements or data generated by testing done in the building, could not remember having read literature concerning asbestos suspension, and had never conducted or seen a study concerning travel of asbestos fibers within an enclosed building. He further testified that he had no personal knowledge of the ventilation system' at Firestone and had not reviewed the system’s equipment specifications. He was unable to quantify the extent of asbestos contamination and exposure. Stanton’s product identification conclusions were derived solely from the testimony of the plaintiffs and the other product identification witnesses rather than from personal knowledge or inspection.
The affidavit of Dr. Robin Kaplan and the deposition testimony of Dr. Alf Fish-beins 8— Although Dr. Kaplan’s testimony was offered in support of the plaintiffs’ motion to reconsider, it did not address the question of the plaintiffs’ exposure to the defendants’ products. This testimony was offered instead to establish the connection between exposure to asbestos and the type of lung changes noted in the plaintiffs.
The deposition testimony of William Pfingsten — Mr. Pfingsten, a Firestone quality inspector from 1951 to 1980, was given responsibility for certifying rubber batches mixed in the compound room and for certifying mill mixes in the mill room. He testified that both soapstone and talc had been used as detackifying agents in the tube room and that he remembered seeing 50 lb. bags of Vermont Talc stored near the duster in the tube room on the 167’ level and in the storage area on the 155’ level. He testified to having seen these bags throughout the course of his employment. Bags of talc, not specifically identified as Vermont Talc, were also seen on the Firestone dock where pigments were unloaded.
VI.
Having detailed all of the evidence bearing on the issue of causation, we must now determine whether the district court erred in concluding that this evidence was insufficient, under Pennsylvania law, to withstand the defendants’ motions for summary judgment.
The district court’s grant of summary judgment was grounded in its conclusion that the plaintiffs had failed, through their own testimony and the testimony of product identification witnesses, to present evidence showing that each plaintiff had inhaled asbestos fibers shed by the particular product of any defendant. The court adhered to the standard articulated in Ecken-rod: a plaintiff’s ability to defeat a motion for summary judgment through circumstantial proof of causation depends upon whether the evidence demonstrates frequent use of the product and “the regularity of plaintiff’s employment in proximity thereto.” 544 A.2d at 53. The district court concluded that the fiber drift theory failed to comport with the Eckenrod holding. “Until the Pennsylvania courts change the law regarding causation in an asbestos case plaintiff’s ... fiber drift theory must fail.” Id.
It is important, at this point, to clarify that the Eckenrod and Samarin standard implicitly takes into account the fact that asbestos fibers may become respirable and “drift.” The plaintiffs meeting the requirements of Eckenrod are bystanders; they can demonstrate no direct contact with products containing asbestos. These plaintiffs are able to survive a motion for summary judgment because where there is evidence showing that a product was used frequently and that a plaintiff worked regularly in proximity to that product, it is reasonable to infer that the plaintiff inhaled drift fibers from that product. When the plaintiffs speak of the “fiber drift theo*376ry,” however, they are not speaking only of the recognition, implicit in Eckenrod, that respirable fibers may be disseminated from a product to the air surrounding that product.
The “fiber drift theory” as it is described by the plaintiffs here takes as its starting point that asbestos fibers may become airborne or re-entrained and thus be carried from their source to other areas. Under this theory, however, both the specific locale of the product’s use and the specific areas of the plaintiff’s employment become irrelevant. The substance of the fiber drift theory is that once an asbestos-containing product can be placed anywhere in the Firestone plant, any plaintiff working at any point within that plant is entitled to have the question of causation submitted to the jury because it is likely, given that fibers can drift, that a given plaintiff was exposed to fibers originating in a particular defendant’s product. When we refer to “fiber drift,” then, we refer to this specific theory and not to the general recognition that fibers may be carried by air currents or other means to locations beyond the point of their release.
In assessing the district court’s award of summary judgment, we consider first only that evidence which does not implicate the fiber drift theory as we have defined it, i.e., only the Eckenrod product identification testimony of the plaintiffs and other plant workers, to determine if the district court may have erred in holding that these plaintiffs did not meet the traditional Eckenrod standard.
As we have discussed, under the strict Eckenrod standard, a plaintiff claiming asbestos-related injury may withstand a defendant’s motion for summary judgment only in those cases in which he is able to raise a reasonable inference that he was exposed to and injured by fibers released by the particular defendant's product. Our review of the record convinces us that, completely without reference to the fiber drift theory, some of the plaintiffs were able to adduce evidence sufficient to establish the requisite Eckenrod showing of regularity, frequency and proximity with regard to some of the defendants’ products and are, therefore, entitled to have the question of causation considered by the jury.
We look first to the plaintiffs’ evidence regarding Allied/Bendix. On the basis of the evidence relating to Allied and Bendix products detailed above, we agree that as to seven of the eight plaintiffs, the causal connection between the injuries alleged and Allied/Bendix products is too tenuous to withstand the motion for summary judgment.
There is, however, one exception to this conclusion. We find that when the evidence bearing on the claims of Plaintiff Stopfel is examined, the Eckenrod standard is satisfied and Stopfel is entitled to have his claims against Allied/Bendix submitted to the jury. From 1965 to 1972, Mr. Stopfel worked in the stock-cutting areas of the Firestone facility. In his deposition, Mr. Stopfel stated, “Basically that was my job, to keep the cutter running.... I ran the mill that supplied the cutters.”
Ronald Leininger, a maintenance mechanic at Firestone from 1967 to 1980, testified that one of his duties involved changing the brake linings on the huge Spadone cutters used in the stock cutting area. Leininger testified that these brake linings contained asbestos and were manufactured by Bendix. To his knowledge, every worn brake lining that he detached from the Spadone cutter was manufactured by Bendix. Finally, plaintiffs’ expert George Stanton offered testimony that asbestos-containing dust is released during the process of braking.
Consequently, with respect to Plaintiff Stopfel, there is record evidence that Allied/Bendix brake linings were used frequently on the two Spadone cutters in the stock-cutting area and that Stopfel worked in proximity to these cutters over a seven year period.9 Thus, the Eckenrod regulari*377ty, frequency, and proximity standard is satisfied without resort to the fiber drift theory and the entry of summary judgment in favor of Allied was error.
Using the Eckenrod standard, we next examine the testimony bearing upon the plaintiffs’ exposure to talc supplied by Vermont Talc, Inc. Vermont Talc’s records indicate that during the years from 1967 to 1980, Vermont Talc supplied at least 1,426,-000 pounds of talc to the Firestone facility. The plaintiffs’ expert mineralogist supplied a report and an affidavit showing that the Vermont Talc product contained .2 to .5 percent asbestiform minerals which would have become airborne and respirable when the talc was used.
William Pfingsten, whose testimony was offered in support of the plaintiffs’ motion to reconsider, testified that he had seen 50 lb. bags of Vermont Talc continuously from 1951 to 1980 in the tube room on the 167’ level and in the storeroom on the 155’ level. Two of the plaintiffs, Beidler and Sgro, testified that they had worked in the tube room where they had been exposed to “soapstone.” Beidler also alleged exposure to talc.
Over a four year period, Plaintiff Beidler worked approximately two weekends per month replacing stands in the tube room and, during the period from 1960 to 1961, he periodically removed insulation there. The air was dusty from soapstone. Plaintiff Sgro worked in the tube room at various jobs over a seven year period. He handled soapstone directly and described the dust in the room generally which resulted from bags of soapstone being dumped into a mixing machine. Sgro did not know whether he had ever handled talc.
Vermont Talc contends that the evidence described was insufficient to withstand the grant of summary judgment because Plaintiffs Beidler and Sgro testified primarily to exposure to soapstone. Vermont Talc stresses the fact that talc and soapstone are distinct products; it does not manufacture soapstone. The record, however, supports the plaintiffs’ contention that there is uncertainty as to whether the workers recognized talc and soapstone as separate products or whether they intended the term “soapstone” to encompass talc. The plaintiffs point to scientific literature in which soapstone has been used as a synonym for “massive talc.”
We believe that Vermont Talc’s reliance on the talc/soapstone distinction will not bear weight in light of the ambiguities in the record; further factual development is required. This is especially true where, if talc and soapstone were viewed as synonymous, Plaintiffs Beidler and Sgro have met the Eckenrod burden with respect to Vermont Talc and are entitled to have the issue of causation submitted to the jury.
For those plaintiffs exposed to talc or soapstone outside of the tube room, crucial product identification testimony linking the particular product to a particular area is missing. These plaintiffs have failed, therefore, to satisfy the requirements of Eckenrod and, under the traditional approach, without reference to fiber drift, the district court’s entry of summary judgment would be proper.
We next turn to the district court’s entry of summary judgment in favor of Defendant John Crane, Inc. While John Crane was awarded summary judgment in all eight cases, its motion for summary judgment encompassed only Plaintiffs Reimert, Moyer, Yourkavitch and Davis. It asks that “the judgment as to plaintiffs Beidler, Sgro, Stopfel and Klein ... be vacated :.. without prejudice.” We will vacate the order of summary judgment in favor of John Crane in those cases. None of the remaining plaintiffs against whom summary judgment was sought is able to satisfy the regularity, frequency and proximity test. Under the Eckenrod standard, then, the district court’s grant of summary judgment in the cases of Reimert, Moyer, Your-kavitch and Davis would be proper.
Finally, we address the district court's conclusions with regard to the products of Eagle-Picher. The district court entered summary judgment in favor of Eagle-Pich*378er as to all eight plaintiffs. Eagle-Picher contends that summary judgment in each case was appropriate and asks that the order be affirmed.
We have examined the deposition testimony of those witnesses identifying Eagle-Picher products in the Firestone plant and conclude that this testimony is insufficient to establish the specific regular use of these products required by Ecken-rod. While several witnesses were able to place Eagle-Picher products in the plant, no witness was able to say how much of the product was used and how many times it was used in a given area. There is no evidence to suggest that any plaintiff was exposed to or worked in the vicinity of any Eagle-Picher product on a regular basis. Under the traditional Eckenrod approach, therefore, we find no error in the district court’s entry of summary judgment in favor of Eagle-Picher.
VII.
Having reviewed all of the available evidence against the Eckenrod standard, we now focus specifically upon the plaintiffs’ fiber drift theory. We do so in order to ascertain whether this theory may be invoked in whole or in part under Pennsylvania law and, if so, whether application of the theory, as expounded by the plaintiffs’ experts, compels a result different from that reached under Eckenrod alone.
The plaintiffs contend that the Eckenrod regularity, frequency and proximity standard regarding proof of causation has no application to those situations in which scientific evidence bearing on causation has been adduced. As we have explained, the plaintiffs argue that once a particular defendant’s product is placed in the Firestone plant during the period of a plaintiff’s employment and expert testimony is offered to show that those products released res-pirable fibers and that air currents could have carried those fibers throughout the plant, a plaintiff has established an inference of exposure sufficient to carry his case to the jury. The plaintiffs argue that where there is expert fiber drift testimony bearing upon causation, Eckenrod is irrelevant and should not be accorded conclusive effect as the law of Pennsylvania.
In the alternative, the plaintiffs argue that if Eckenrod is held to be applicable even where the fiber drift theory has been invoked, the totality of the evidence offered by the plaintiffs is sufficient to satisfy the Eckenrod regularity, frequency and proximity test.
Our first focus in evaluating the plaintiffs’ fiber drift arguments must be upon our role with respect to Pennsylvania law. As a basic premise, federal courts sitting in diversity cases are required to apply the substantive law of the state whose laws govern the action. Erie Railroad Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938). In cases where the state’s highest court has not considered the precise question to be answered, the federal court is called upon to predict how the state court would resolve the issue should it be called upon to do so. Commission v. Estate of Bosch, 387 U.S. 456, 465, 87 S.Ct. 1776, 1782, 18 L.Ed.2d 886 (1967); Pennsylvania Glass Sand Corp. v. Caterpillar Tractor Co., 652 F.2d 1165, 1167 (3d Cir.1981).
In making such predictions we recognize that “[t]he state’s highest authority is the best authority on its own law” and that “if there be no decision by that Court then federal authorities must apply what they find to be the state law after giving ‘proper regard’ to the relevant rulings of other courts of the state.” Id. Our role is not to form or create state law but to decide the case as we believe it would have been decided by the state’s highest court had the case arisen in the state court system. Becker v. Interstate Properties, 569 F.2d 1203, 1205 n. 5 (3d Cir.1977), cert. denied, 436 U.S. 906, 98 S.Ct. 2237, 56 L.Ed.2d 404 (1978).
While it is clear that the Pennsylvania Supreme Court has not considered a case on all fours with this one, it has definitively established what a plaintiff is required to prove in order to prevail in a products liability action. In particular, the court has required that a plaintiff demon*379strate that his harm was proximately caused by the product of a particular defendant.10 Berkebile, 337 A.2d at 893. Applying this general principle, Pennsylvania’s trial and intermediate appellate courts, and federal courts applying Pennsylvania law, have held consistently that in asbestos-related personal injury cases, general products liability principles apply; the plaintiff is required to establish that the product of a specific manufacturer caused the injury alleged. See, e.g., DiSantis v. Abex Corp., C.A. No. 87-0515, 1989 WL 150548 (E.D.Pa. Dec. 8, 1989) (in asbestos case defendant cannot be held liable to any party unless it is proven that the plaintiff was exposed to asbestos from defendant’s product and that such exposure was the proximate cause of the plaintiff’s injury); Kittrick v. GAF Corp., 125 F.R.D. 103 (M.D.Pa.1989) (citing Eckenrod and holding that in order to avoid summary judgment third-party plaintiff was required to offer evidence that plaintiff was exposed to products of third-party defendant either by working with them himself or working in the vicinity of them); Richards v. Raymark Industries, 660 F.Supp. 599 (E.D.Pa.1987) (in order to recover plaintiff must show that he was exposed to defendant's products, either by working with or in the vicinity of them); Vigiolto v. Johns-Manville Corp., 643 F.Supp. 1454 (W.D.Pa.1986) aff'd mem. 826 F.2d 1058 (3d Cir.1987) (in asbestos case, proof that a specific defendant caused the plaintiff’s harm is required); Pongrac v. Consolidated Rail Corp., 632 F.Supp. at 126 (summary judgment granted where plaintiff failed to establish exposure to defendant’s product); Samarin v. GAF Corp., 574 A.2d at 71 (reiterating application of Eckenrod standard to asbestos cases in Pennsylvania).
In examining the available case law analyzing the plaintiffs’ burden of proof, we see no indication that the courts of Pennsylvania or those predicting the course of Pennsylvania law are prepared to retreat from what is admittedly a stringent standard of proof, even in those cases where the circumstances surrounding the alleged injury make it difficult for a plaintiff to identify the manufacturers of the product responsible for the injury. In Burnside v. Abbott Laboratories, 351 Pa.Super. 264, 505 A.2d 973 (1985), the Superior Court of Pennsylvania declined to adopt the market share, enterprise or concerted action theories of liability in a products liability case involving exposure to DES. “A cause of action does not exist unless the plaintiff can identify the manufacturer ... which caused the injury.” 505 A.2d at 983.
Considering the same liability theories in an asbestos exposure case, the United States District Court for the Western District of Pennsylvania has twice predicted that the Pennsylvania Supreme Court would not liberalize its approach to causation in cases involving asbestos-containing products. In Prelick v. Johns-Manville Corporation, 531 F.Supp. 96, 98 (W.D.Pa.1982), the district court held that where the plaintiff was able to identify at least one supplier or manufacturer whose product caused his injury, “enterprise” liability had no application. Enterprise liability was again rejected in Vigiolto v. Johns-Manville Corp., 643 F.Supp. 1454, 1460-65 (W.D.Pa.1986). The district court adopted the language of the Court of Appeals for the Eleventh Circuit in Blackston v. Shook & Fletcher Insulation Co., 764 F.2d at 1483, in finding “significant policy reasons for refusing to impose market share liability:”
First, elimination of a causation requirement would render every manufacturer an insurer not only of its own products, but also of all generically similar products manufactured by its competitors. Second, expanding culpability of asbestos manufacturers could reduce the ability to spread losses by insurance and otherwise distribute risk. Third, application of such a novel theory of causation would raise serious questions of fairness due to the fact that different manufacturers’ asbestos products differ in degrees of *380harmfulness. Thus, even if this Court were not bound ... to require proof of exposure to a particular defendant’s products to establish proximate cause, significant policy reasons favor retention of proximate cause as an essential element of a cause of action in asbestos litigation.
643 F.Supp. at 1464 (citations omitted). The district court noted that courts in other jurisdictions had declined to impose market share liability in asbestos litigation for similar policy reasons and predicted that “the Pennsylvania Supreme Court, if presented with a similar issue, would find that asbestos litigation is an inappropriate' context in which to extend the market share theory of liability."11 Id.
The prediction of law set forth in Vigiol-to is consistent with the approach taken by a majority of state and federal courts in considering arguments favoring a more liberal approach to proof of causation in asbestos exposure cases. See, e.g., Blackston, 764 F.2d at 1480 (policy reasons favor retention of proximate cause in Georgia law); Thompson v. Johns-Manville Sales Corp., 714 F.2d 581, 583 (5th Cir.1983), cert. denied, 465 U.S. 1102, 104 S.Ct. 1598, 80 L.Ed.2d 129 (1984) (supposed general tendency on the part of Louisiana courts to expand liability of manufacturers was held insufficient to warrant district court’s adopting radical departure from traditional theories of tort liability represented by enterprise and market share liability); Sholtis v. American Cyanamid Co., 568 A.2d at 1196 (market share theory rejected as impractical where there was possible exposure to an unknown mix of products over an extended period); Goldman v. Johns-Manville Sales Corp., 33 Oh.St.3d 40, 514 N.E.2d 691 (1987) (asbestos-containing products do not create similar risks of harm required for application of alternative liability theory); Case v. Fibreboard Corp., 743 P.2d 1062 (Okla.1987) (Oklahoma would not recognize collective liability as a theory of relief in asbestos litigation where plaintiff was unable to identify specific tort-fea-sor); Celotex v. Copeland, 471 So.2d 533 (Fla.1985) (market share theory inappropriate where asbestos worker was able to identify majority of manufacturers supplying asbestos); Gaulding v. Celotex Corp., 772 S.W.2d 66 (Tex.1989) (where all possible tortfeasors not joined as defendants, alternative liability theory unavailable).
Based on the foregoing analysis, we can find no basis for predicting that the Supreme Court of Pennsylvania would disavow the Eckenrod regularity, frequency, and proximity standard in this case simply because this case involves expert scientific evidence rather than traditional co-worker testimony. The fiber drift theory cannot stand alone; it must be supported by evidence showing the frequency of a product’s use and the regularity of the plaintiff’s employment in an area into which there is a reasonable probability that the fibers drifted. In reaching this conclusion we are mindful that in predicting the parameters of Pennsylvania law we “must be sensitive to the doctrinal trends of the state ... and the policies which inform the adjudications by the state courts.” Becker v. Interstate Properties, 569 F.2d at 1206. We are unable, however, to ascertain any trend in the law of Pennsylvania toward an expansion of liability on the part of manufacturers of asbestos through relaxation of the burden of proof relative to causation.
In an effort to convince us that the Pennsylvania Supreme Court would not follow Eckenrod given the nature of the circumstantial evidence presented here, the plaintiffs place principal reliance upon the decision of the Washington state intermediate appellate court in Lockwood v. AC & S, Inc., 44 Wash.App. 330, 722 P.2d 826 (1986), aff'd, 109 Wash.2d 235, 744 P.2d 605 (1987) (en banc). In Lockwood, the plaintiffs filed suit claiming injury as a result of shipyard exposure to asbestos-containing products. Defendant Raymark appealed the jury verdict in plaintiffs' favor contending, among other things, that the plaintiffs *381had failed to present evidence of causation sufficient to carry the question of liability to the jury.
The causation evidence presented in Lockwood consisted of expert testimony tending to show that asbestos exposure causes asbestosis and that once asbestos dust is released, it can remain in the air and be carried on air currents through a work area over a period of time. Additional testimony established that the defendants’ products were present in the shipyards during the plaintiffs employment there. The court summarized Raymark’s challenge to the adequacy of the causation evidence as follows:
Raymark argues that the evidence presented by Lockwood was insufficient to establish that exposure to its product proximately caused his injury and, therefore, the question of Raymark’s liability should not have been presented to the jury. Raymark emphasizes that Lockwood did not personally handle asbestos products in his work, and that his primary exposure to asbestos occurred when asbestos was torn from vessels during rigging. It observes that there was no specific, direct evidence that asbestos torn by riggers was asbestos cloth, let alone cloth made by Raymark. Raymark also points out that the identification of its cloth at trial related to the application of asbestos rather than the tearing out of old asbestos. In addition, other asbestos cloth manufacturers served shipyards where Lockwood worked. Raymark also relies on testimony by one of its officials that asbestos cloth is cleaner than other types of asbestos products such as cement and block products. Raymark further emphasizes that Lockwood himself never personally identified its product as present during ship repair. In addition, there is no direct evidence that Lockwood worked with or near Raymark cloth on the George Washington or that Ray-mark cloth was torn from that vessel.
744 P.2d at 611. In upholding the adequacy of the causation evidence despite the challenges raised, the Washington court recognized that traditional products liability theory requires that the plaintiff establish a “reasonable connection” between the alleged injury and the defendant’s asbestos-containing product. The court, however, expressed a willingness to depart from traditional precepts noting that it had “eased the strict requirements of the traditional approach where there are unusual problems involved in product identification.” Id. at 612 n. 6. While that court had previously adopted a market share theory of liability in a DES case, Martin v. Abbott Laboratories, 102 Wash.2d 581, 689 P.2d 368 (1984), it specifically disavowed any reliance on alternative liability theories in affirming the jury award in Lockwood. “[The] plaintiff in the present case presented evidence as to the identity of manufacturers of the products at his workplace, which in combination with other evidence, satisfies the requirements of traditional proximate causation.” Id.
The court in Lockwood recognized that it would be extremely difficult to determine whether a plaintiff’s injury were caused by a particular defendant’s product. Offering guidance to other courts facing asbestos-related causation issues, the court directed that several factors be considered: the plaintiff’s proximity to the asbestos product at the time of exposure; the expanse of the workplace where fibers were released; the length of a plaintiff’s exposure; the types of asbestos products involved and their tendency to release asbestos fibers into the air; and the evidence presented on medical causation. “Ultimately the sufficiency of the evidence of causation will depend on the unique circumstances of each case.” 744 P.2d at 613.
Despite the Washington court’s articulation of traditional-sounding factors to be considered in determining the sufficiency of a plaintiff’s causation evidence, the court reached a non-traditional result in finding that the mere presence of the defendant’s product in the workplace coupled with “fiber drift” evidence was sufficient to justify submission of the issue of causation to the jury.
In our view, Lockwood adopts a causation standard much more expansive than *382the Pennsylvania standard introduced in Eckenrod. We have located no other case adopting this relaxed causation standard and note that at least one court has characterized the Lockwood evidentiary nexus as “extremely attenuated.” Menne v. Celotex Corp., 861 F.2d 1453, 1462 at n. 12 (10th Cir.1988). As we have noted, the vast majority of the cases addressing circumstantial evidence in the context of asbestos litigation requires more, i.e., some concrete demonstration of a link between the alleged injury and a defendant’s particular asbestos-containing product.
While we understand the plaintiffs’ emphasis on Lockwood, we find no ground whatever for predicting that the Supreme Court of Pennsylvania in evaluating this case would repudiate the nexus requirement of Eckenrod and adopt the Washington Supreme Court’s reasoning.
The plaintiffs purport to find support for an abandonment of Eckenrod in two federal court decisions. They cite our decision in Rocco v. Johns-Manville Corp., 754 F.2d 110 (3d Cir.1985), and the decision of the district court in Wible v. Keene Corp., No. 86-4451, 1987 WL 15833 (E.D.Pa.Aug. 19, 1987), for the proposition that these courts “have not taken the rigid approach suggested by Eckenrod.” We do not read either of these cases to support a move away from Eckenrod.
While we held in Rocco that the placement of a defendant’s asbestos-containing product in the plaintiff’s workplace was sufficient evidence of causation, we acknowledged that the evidence was “not overwhelming.” 754 F.2d at 113. Moreover, the Rocco decision pre-dated Ecken-rod and thus its continued validity in the circumstances presented in this case is questionable.
Wible is equally unavailing. In that case, witnesses were able to identify the specific manufacturer of the asbestos product alleged to have caused injury and testified to its use throughout the workplace. Testimony also established that the plaintiff had been employed in the plant storeroom where all products used in the plant were brought for distribution. The Wible court concluded that the evidence presented was sufficient to support the inference that the plaintiff worked in the vicinity of the product in question on a regular basis. This case is, in our view, completely consistent with the Eckenrod test and in no way suggests that the law of Pennsylvania is evolving away from that standard.
The plaintiffs’ final argument in favor of a repudiation of Eckenrod is grounded purely in policy considerations; they cite the injustice done to bystanders by adherence to Eckenrod's strict requirements of proof. As we have stated, even were we to disagree with Eckenrod, our role in this matter is limited. Our role is to predict rather than to shape the law of Pennsylvania. We do not write on a clean slate and are bound by the policy implications of Pennsylvania law as we find it to exist.
In sum, we reject the plaintiffs’ contention that the district court erred in finding the law of Eckenrod controlling in this case; we predict that the Pennsylvania Supreme court would find that the Eckenrod requirements of regularity, frequency and proximity apply with equal force in cases involving proof of causation through the use of expert scientific testimony as in cases relying upon more traditional circumstantial evidence. Under Pennsylvania law, the fiber drift theory, as defined by the plaintiffs, is insufficient to create a jury question on the issue of causation.
VIII.
Having predicted the vitality of the Eckenrod causation standard in assessing the evidence in this case, we move to our final determination: whether the district court erred in concluding that the fiber drift testimony supplied by the plaintiffs’ experts was inadequate under Eckenrod.12
*383Of the three Eckenrod requirements, regularity, frequency and proximity, the fiber drift theory bears on proximity alone; it sheds no light on how often a particular product was used, nor does it establish how often a particular worker was present in an area into which fibers may have drifted. It is only when regularity and frequency are established that testimony implicating fiber drift and proximity becomes relevant.
In this case, the district court concluded that none of the plaintiffs had established regularity and frequency and that, therefore, expert testimony that fibers can drift from one location to another was not probative of the plaintiffs’ exposure to the defendants’ products.13 This conclusion finds support in a decision of the Court of Appeals for the Tenth Circuit. That court held that expert evidence offered to prove that asbestos-containing products caused cancer was insufficient as a matter of law to establish that a plaintiff’s exposure to the defendants’ particular products was “sufficiently sustained, frequent or intense” to constitute either a cause in fact or a legal cause of injury. Menne v. Celotex Corp., 861 F.2d at 1463-64. We find that the court’s language there applies in this case as well: “From the circumstantial evidence, there [was] no way of ascertaining the regularity or frequency of [the plaintiffs’] exposure to ... asbestos dust from a given defendants’ products. Such exposure might have been frequent or sustained or it might have been sporadic and short. On this matter the jury could only speculate.” Id. at 1465.
Our reading of Pennsylvania case law and the law in other jurisdictions bearing on causation in asbestos cases convinces us that competent factually based expert testimony regarding fiber drift may be offered by a plaintiff where it rests on a properly laid foundation showing the frequency with which a defendant’s product was used, the area in which it was used and the regularity of a plaintiff’s employment within a zone covered by the reach of fiber drift.14 While the fiber drift theory may be useful in extending the zone of proximity, it cannot, standing alone, allow a plaintiff to circumvent the other requirements of Eckenrod.
While we are able to validate the approach of the district court, we cannot validate its conclusions in every respect. Our review of the record convinces us that the Eckenrod foundation requirements were satisfied in the cases of Plaintiff Stopfel’s exposure to Allied/Bendix brake linings and may, depending upon factual developments regarding the distinction between talc and soapstone, have been satisfied as to Plaintiffs Sgro’s and Beidler’s exposure to Vermont Talc. Competent factually supported expert testimony regarding fiber drift and possible exposure from these products in the established areas of use is appropriate.
IX.
Because we conclude that Eckenrod is the law of Pennsylvania and controls even *384in cases where expert testimony is introduced on the issue of causation, we will affirm the district court’s grant of summary judgment in favor of the defendant Eagle-Picher with respect to all plaintiffs.
We will affirm the district court’s grant of summary judgment with respect to John Crane, Inc. in the cases of Plaintiffs Rei-mert, Yourkavitch, Moyer and Davis. Consistent with the request of John Crane, Inc., we will vacate the district court’s grant of summary judgment in the cases of Beidler, Klein, Sgro, and Stopfel and will remand for further proceedings consistent with this opinion.
With respect to Allied Signal, Inc. we will vacate the district court’s grant of summary judgment as to all plaintiffs, each of whom worked on the 180' level during the relevant period, pending evaluation of the experts’ fiber drift testimony in light of evidence showing that Allied/Bendix brake linings were used on the Spadone cutter on the 180' level regularly between the years 1967 and 1980. With respect to Vermont Talc, Inc. we will affirm the district court’s grant of summary judgment as to Plaintiffs Davis, Moyer, Reimert, Yourkavitch, Kline and Stopfel as none of these plaintiffs worked in the enclosed tube room during the period when Vermont Talc products were placed in that room. We will vacate the entry of summary judgment as to Plaintiffs Beidler and Sgro, as a result of their employment in the tube room, and will remand these cases for further factual development on the issue of the talc/soapstone distinction.
. The eight worker-plaintiffs and their wives were divided into two trial groupings: Plaintiffs Reimert, Yourkavitch, Moyer and Davis constituted the first group, while the second group was comprised of Beidler, Sgro, Stopfel and Kline.
. Although Defendants John Crane and Eagle Picher did not file motions for summary judgment in the cases involving Plaintiffs Sgro, Stop-fel, Beidler and Kline, the district court entered summary judgment in those cases. The plaintiffs note that summary judgment was entered in cases where no motion was filed. They did not, however, challenge this entry below and have not made this issue a basis for appeal. John Crane, Inc. acknowledges that, as to it, the entry of summary judgment was an administrative error and asks that the order be vacated.
. The question was certified as follows:
Whether under the decision of the Pennsylvania Superior Court in Eckenrod v. GAF Corporation and this court in Pongrac v. Consolidated Rail Corporation the plaintiffs can, through their own testimony or the testimony of product identification witnesses demonstrate that each inhaled asbestos fibers shed by the asbestos-containing products of each defendant, or whether, under Pennsylvania law, plaintiffs may substitute proof of their exposure to defendants’ products by means of the ambient air theory? (Citations omitted.)
. E.g., Bracey v. Keene, C.A. No. 86-6830, 1989 WL 156380 (E.D.Pa.Dec. 18, 1989) (adopts Eck-enrod test); Harkovich v. Keene, 745 F.Supp. 1062, 1064 (M.D.Pa.1989) (evidence that demonstrates the general or possible presence of a particular asbestos product at a particular job site is not enough to defeat summary judgment); McClaskey v. Anchor Packing Co., C.A. No. 86-7599 1989 WL 114346 (E.D.Pa.Sept. 28, 1989) (under Eckenrod, a plaintiff may withstand summary judgment on causation grounds only by offering "competent evidence of the length or nature of plaintiff-husband’s contact with defendant’s product”); Williams v. Celotex Corp., C.A. No. 85-550-JJF (D.Del. Jan. 24, 1989) (under Pennsylvania law, as articulated in Eckenrod, defendant entitled to directed verdict where plaintiff failed to adduce evidence that plaintiff "was exposed to the asbestos-containing products of this specific defendant ... on a regular basis over a sufficient period of time”), aff'd mem. sub nom. Williams v. AC & S, 882 F.2d 513 (3d Cir.1989); Pongrac v. Consolidated Rail Corp., 632 F.Supp. 126 (E.D.Pa.1985) (plaintiff showed no evidence that he had worked in the vicinity of asbestos products.)
. It is apparent that the Pennsylvania Superior Court took the requirements of Eckenrod from Schmidt v. Johns-Manville Corp., C.A. No. 80-3339 (D.Md. Nov. 30, 1982). See also Blackston v. Shook & Fletcher Insulation Co., 764 F.2d 1480, 1481 (11th Cir.1985) (plaintiff must show that he was exposed to defendant’s asbestos-containing product by working with or in close proximity to the product); Roberts v. Owens-Corning Fiberglas Corp., 726 F.Supp. 172, 174 (W.D.Mich.1989) ("plaintiff must establish that the manufacturer’s asbestos product was used at the specific site within the workplace where he worked”); Slaughter v. Southern Talc Co., Civil No. W-87-CA-060, slip op. at 6 (W.D.Tex. Mar. 28, 1990) (adopting Lohrmann test under Texas law citing Eckenrod; Grubb v. Norfolk & W. Ry. Co., C.A. No. 85-0751-R, slip op. at 2 (W.D.Va. Nov. 10, 1986) (to withstand summary judgment on causation ground, "[p]laintiff must show that he was exposed to the product in question on a regular basis over an extended period of time.”); Sholtis v. American Cyanamid Co., 238 N.J.Super, 9, 568 A.2d 1196, 1207 (App.Div.1989) (adopting "well-reasoned" ” ‘frequency, regularity and proximity’ ” test of causation in asbestos litigation); Little v. Techbestos, Inc., No. A-4104-87T2, slip op. at 7-10 (N.J.App.Div. Jan. 8, 1990) (same); O'Connor v. Raymark Indus., Inc., 401 Mass. 586, 518 N.E.2d 510, 511 (1988) (jury instruction that proof of causation requires "evidence of some exposure ... on a regular basis over some period of time where [the worker] was actually working with the [asbestos] product himself or in proximity to where others were working with the product”); Zimmer v. Celotex Corp., 192 Ill.App.3d 1088, 140 Ill.Dec. *369230, 233, 549 N.E.2d 881, 884 (1989) ("plaintiff must show that a particular defendant’s product was used at the job site and that the plaintiff was in proximity to that product at the-time it was being used”).
. John Crane reiterates that it did not move for summary judgment in the case of Plaintiffs Bei-dler, Sgro, Stopfel and Kline.
. The testimony of William Pfingsten, a Firestone quality inspector during the period from 1951 to 1980, recalled seeing 50 lb. bags in the tube room with the words “Vermont Talc" on them throughout the course of his employment. Sgro and Reimert worked in this area but claimed exposure to soapstone rather than talc. Pfingsten’s testimony was not submitted for the court’s consideration until the time of the motion for reconsideration.
. The testimony of Dr. Fishbein, an “occupational/environmental medical expert witness,” is not included in the record nor is its substance summarized or referred to in the plaintiffs’ brief. The parties agree, however, that this testimony was not product-specific.
. While Plaintiff Moyer also worked in the stock-cutting room, his employment there did not coincide with Leininger’s. There is, there*377fore, no link to Allied/Bendix products in this instance.
. While the state’s highest court has yet to review this requirement in a case of asbestos-related injury, it is, in our view, significant that the Pennsylvania Supreme Court refused alloca-tur in both Eckenrod and Samarin.
. The court ruled that alternative liability and enterprise liability were unavailable on the facts presented.
. At this point it is important to note that Pennsylvania law does not categorically preclude the introduction of expert testimony on the issue of causation in a products liability action, nor do we believe that the district court so held. See Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280 (1978). We read the district court's opinion regarding exclusion of the scientific tes*383timony as having been based instead upon the lack of a factual foundation strong enough to support the inference of causation. We note that an inference based upon a speculation or conjecture does not create a material factual dispute sufficient to defeat entry of summary judgment. Blackston v. Shook & Fletcher Insulation Co., 764 F.2d 1480, 1482 (11th Cir.1985).
. Courts utilizing the "frequency, regularity, and proximity" test have applied the test in cases where the evidence of exposure is based upon drifting asbestos fibers. See Brady v. Raymark Indus., Inc., 896 F.2d 558 (11th Cir.1990) (per curiam) (lay testimony on theory that dust containing asbestos fibers traveled throughout plant); Blackston v. Shook & Fletcher Insulation Co., 764 F.2d at 1480 (affidavit from counsel with article submitted as a "learned treatise" and co-worker testimony on the nature of asbestos dust); Slaughter v. Southern Talc Co., Cause No. W-87-CA-060, (W.D.Tex. Mar. 28, 1990) (expert opinion on re-entrainment and product identification witnesses' testimony); and Williams v. Celotex Corp., C.A. No. 85-550 (D.Del. Jan. 24, 1989) (court disregarded fiber drift testimony based on Eckenrod), aff’d sub nom. Williams v. A.C. & S, Inc., 882 F.2d 513 (3d Cir.1989).
. While we are aware that the defendants have vigorously challenged the probative value of the scientific experts’ testimony, we believe that our jurisprudence is best served by having the district court evaluate these challenges under Rules 703 and 403 of the Federal Rules of Evidence. See Viterbo v. Dow Chemical Co., 826 F.2d 420 (5th Cir.1987).