ROVNER

Richard L. Carlson (argued), Attorney, Hunegs, LeNeave & Kvas, P.A., Minneapolis, MN, Neal C. Zazove, Attorney, Zazove & Associates, Chicago, IL, for Plaintiff–Appellant. Marivel Montes (argued), Attorney, Metra, Chicago Regional Transportation Authority, Chicago, IL, for Defendant–Appellee.
Before KANNE, ROVNER, and WILLIAMS, Circuit Judges.

ROVNER, Circuit Judge.

On October 8, 2007, Renardo Lynch was injured while working at a jobsite as a mechanic for Metropolitan Rail (Metra), when the top rail of a chain-link fence he was installing fell and struck him on the back of his neck and shoulders. Lynch filed a complaint under the Federal Employers' Liability Act (FELA) seeking damages against Northeast Regional Commuter Railroad Corporation, doing business as Metra, for the injuries he sustained that day. The district court granted summary judgment in favor of Metra, and Lynch appeals.

I.

Lynch was hired by Metra in 1987 to work in the track department, but moved to the Bridges and Building (“B & B”) department where he held a number of positions. At the time of the injury, he was working as a B & B mechanic. The duties of a mechanic included: installing fences, doors and windows; painting; brickwork; installing pedestrian road crossings at depots; upkeep of depots; and maintaining Metra bridges and buildings. Although Metra provided training regarding some of those duties, no training was provided regarding the installation of fencing. Mechanics learned how to install fences from working with peers on the jobsites.

The installation of fences was a routine part of a mechanic's job in that such work was done several times per month, and it occurred in distinct phases over multiple days. In the first stage, a work crew would dig holes about three feet deep and set vertical fence posts in cement. Those posts, called terminal or end posts, measured approximately 3 inches in width. The cement was then allowed to cure for 1–2 days.

In the next phase, the top rails of the fence were installed. Those rails were secured to the fence posts by means of brackets with attached cups that were tightened around the fence post. The cups were recessed at least one and a half inches so the top rail could be placed in the sleeve of the cup and secured. The top rail was first cut to the proper dimension to fit from one cup end to another, and the cups were loosened to position the top rail in place and then tightened to secure it. According to the deposition testimony of crew members who regularly installed fencing, once the top rail is secured and the brackets tightened the top rail should not be able to slip out of the cups.

After the top rail is installed, the fabric or chain link is put in place and secured to the skeleton—the rail and posts. Lynch was engaged in this task at the time of the injury.

On the day of the incident, Lynch and the other members of his work crew reported to Metra's Western Avenue facility at 6:00 in the morning. The work crew that day included the foreman Brad Clark, assistant foreman Trancito Reyes, B & B mechanics Ivory Scott and Kurtis Otero, and Nathan Fullbright. The foreman briefed the crew as to their tasks for the day, and they proceeded to the Western Avenue depot with the necessary supplies. It is undisputed that the fence posts had previously been installed at that site, and the evidence is unclear as to whether the top rails were in place or whether Lynch's work crew installed them that day. Lynch did not believe that he helped install the top rail that day. The fence was located on a hill, which was described as steep, and there was a drop in elevation between the two fence posts.

At the time of the incident, Lynch and Otero were installing the fabric or mesh part of the fence and were on their knees next to each other tightening brackets at the bottom of the fence post. The top rail fell, hitting Lynch across the back of his neck and shoulders and causing him to sprawl “flat face down.” Lynch was uncertain as to whether he lost consciousness, but he was dazed. He ended up missing work for approximately 28–30 days following the injury.

Lynch and Otero both maintained that they were not pulling on any portion of the posts or top rail at the time the rail dropped, and that they did not believe any actions on their part contributed to its fall. Metra has acknowledged that there was nothing Lynch or his co-workers did to cause the pole to fall. See Metra's Rule 56.1 Statement of Uncontested Facts at 3, no. 29. Metra also admitted that the employees are responsible for inspecting the work being done. Id. at 4, no. 38.

II.

The district court recognized that under FELA, 45 U.S.C. § 51 et seq., an employee will survive summary judgment if the evidence justifies with reason the conclusion that the employer's negligence played any part in producing the injury. Dist. Ct. Op. at 4. According to the district court, that means that under FELA an employer is liable for its negligence even if the injured worker is even more negligent, but it does not stand for the proposition that a plaintiff who “fails to produce even the slightest evidence of negligence” is entitled to a jury trial. Id. The court then considered the evidence produced by Lynch to establish the elements of negligence. Although Metra had moved for summary judgment on the basis that Lynch failed to demonstrate a breach of due care, the court assumed that Lynch had in fact presented sufficient evidence of a breach of its duty to provide a reasonably safe workplace with proper training. Id. at 5–6. Instead, the court granted summary judgment for Metra on the issue of causation, holding that Lynch's theory that the top rail was cut too short or improperly installed rested on speculation not facts. Id. at 6. In particular, the court emphasized the failure of Lynch to introduce evidence of the measurement of the top rail and the distance between the fence posts, or the grade of the hill. Id. at 7–8. The court dismissed the testimony of coworker Otero that a top rail should not come loose if cut and secured properly, declaring that Otero was a fact witness not an expert, and that no expert testimony was provided. Id. at 7. The court held that the causation standard under FELA was not so lax as to allow a plaintiff to proceed on nothing more than rank speculation, and granted summary judgment to Metra on that basis. Id. at 8.

In addressing the lack of evidence presented by Lynch regarding causation, the district court noted that it could be related to the failure of Metra to argue for summary judgment on that issue:

To be fair, plaintiff's brevity on this issue is likely the result of defendant's curious failure to argue that it is entitled to summary judgment based on the absence of any evidence of causation, instead focusing on the argument that the evidence does not support a breach of due care.
Id. at 6, n. 3. The court nevertheless concluded that Metra had “carried its initial burden under Rule 56, by identifying record evidence to support its statement that ‘it is unknown why the pole fell,’ and is entitled to summary judgment based on plaintiff's failure to dispute that evidence with ‘specific facts showing there is a genuine issue for trial.’ ” Id.

III.

On appeal, Lynch argues that the district court erred in granting summary judgment on a basis not asserted by Metra without providing it an opportunity to respond. Pursuant to Federal Rule of Civil Procedure 56(a), “a party may move for summary judgment, identifying each claim or defense—or the part of each claim or defense—on which summary judgment is sought.” In this case, Metra moved for summary judgment on the ground that Lynch had failed to demonstrate negligence, but specifically that he had not demonstrated a breach of duty because Metra lacked actual or constructive notice of a defect that caused Lynch's injuries. The court chose, however, to grant summary judgment on a different ground not argued by Metra, that of causation.

Metra does not contend on appeal that it briefed the causation issue before the district court. Instead, it points out that under Federal Rule of Civil Procedure 56(f), a district court may grant summary judgment on a ground not raised by the moving party. A look at the actual language of Rule 56(f) clarifies the circumstances under which the court may so act:

After giving notice and a reasonable time to respond, the court may:

(1) grant summary judgment for a nonmovant;

(2) grant the motion on grounds not raised by a party; or

(3) consider summary judgment on its own after identifying for the parties material facts that may not be genuinely in dispute.
[emphasis added]. Rule 56(f) thus allows a court to grant summary judgment on grounds not raised by a party only after providing notice and a reasonable time to respond. There is no indication that such notice and time to respond was provided in this case. Lynch asserts on appeal that, given an opportunity to respond, he would have presented medical evidence linking his injury to the impact caused by the top rail. That evidence, however, would not have addressed the court's concern that Lynch had failed to demonstrate that Metra's breach of a duty caused the top rail to fall. It is unclear whether Lynch would have presented evidence relating to that issue, such that the failure to provide notice and time to respond would have adversely impacted him. Ultimately, we need not address this Rule 56(f) issue, because there was no need for Lynch to provide any further response; we hold below that the record contains sufficient evidence of causation and therefore the court improperly granted summary judgment on the merits.

IV.

FELA was enacted in response to the dangers inherent in working for the railroad and the high rate of injuries among railroad employees. See Consolidated Rail Corp. v. Gottshall, 512 U.S. 532, 542–43, 114 S.Ct. 2396, 129 L.Ed.2d 427 (1994). It establishes a standard for employer liability that is more lax than common law negligence standards, and eliminates a number of traditional defenses such as contributory negligence, the fellow-servant rule, and assumption of risk. Williams v. National R.R. Passenger Corp., 161 F.3d 1059, 1061 (7th Cir.1998); Gottshall, 512 U.S. at 542–43, 114 S.Ct. 2396. Under FELA, railroads are liable if carrier negligence played any part, even the slightest, in producing the injury. CSX Transportation, Inc. v. McBride, ––– U.S. ––––, 131 S.Ct. 2630, 2634, 180 L.Ed.2d 637 (2011); DeBiasio v. Illinois Central R.R., 52 F.3d 678, 685 (7th Cir.1995). A plaintiff's burden under FELA is thus significantly lighter than in an ordinary negligence action. Green v. CSX Transportation, Inc., 414 F.3d 758, 766 (7th Cir.2005). A jury verdict in a FELA action can be set aside only if there is a complete absence of probative facts to support the jury's conclusion. DeBiasio, 52 F.3d at 685;Hines v. Consolidated Rail Corp., 926 F.2d 262, 268 (3d Cir.1991) (holding that a “ ‘trial court is justified in withdrawing ... issue[s] from the jury's consideration only in those extremely rare instances where there is a zero probability either of employer negligence or that any such negligence contributed to the injury of an employee.’ ”).

FELA imposes strict liability on railroad carriers who violate certain safety statutes, but none of those statutes are implicated here. See McBride, 131 S.Ct. at 2643 n. 12;McGinn v. Burlington Northern R. Co., 102 F.3d 295, 298–99 (7th Cir.1996); Granfield v. CSX Transportation, Inc., 597 F.3d 474, 480 (1st Cir.2010); Phillips v. CSX Transportation, Inc., 190 F.3d 285, 288 (4th Cir.1999). Therefore in order to survive summary judgment, Lynch had to allege evidence creating a genuine issue of fact on the elements of negligence including duty, breach, foreseeability, and causation. Green, 414 F.3d at 766. Before the district court, Metra argued that it was entitled to summary judgment because Lynch failed to present evidence that Metra was on notice of any unsafe condition and did not create a genuine issue of fact that Metra breached its duty to provide a safe workplace. The district court assumed that Lynch met the element of breach of duty, but granted summary judgment on the ground that Lynch did not raise a genuine issue of fact as to the element of causation. On appeal, the only issue presented to us is whether the district court properly granted summary judgment based on Lynch's failure to raise a genuine issue of fact as to causation. Metra does not argue on appeal that summary judgment should be upheld based on other grounds such as duty, breach or foreseeability, and therefore we can limit our analysis to whether the district court properly resolved the causation issue.

A.

The district court began by correctly stating that an employee in a FELA action is “entitled to a jury if ‘the proofs justify with reason the conclusion that employer negligence played any part, even the slightest, in producing the injury,’ ” a standard set forth by the Supreme Court in Rogers v. Missouri Pacific R.R. Co., 352 U.S. 500, 506, 77 S.Ct. 443, 1 L.Ed.2d 493 (1957). Dist. Ct. Memorandum Opinion and Order at 4, quoting Harbin v. Burlington Northern R.R. Co., 921 F.2d 129, 131 (7th Cir.1990) and Rogers (emphasis in Harbin ). The district court then concluded, however, that “the import of this principle is merely that under the FELA, an employer will be liable for its negligence even if the injured worker was even more negligent,” and declared that it does not stand for the proposition that a plaintiff who fails to produce even the slightest evidence of negligence is entitled to proceed to a jury trial. Id. That characterization of the FELA standard is troublesome. The Supreme Court in McBride, decided after the district court issued its opinion, rejected the notion that the “any part ... in causing the injury” language concerned only division of responsibility among multiple actors, and not causation generally. 131 S.Ct. at 2638 n. 2. In that case, CSX argued that the relaxed FELA standard displaced only common law restrictions on recovery for injuries involving contributory negligence, and did not address the requisite directness of a cause. Id. at 2637. The Court held that Rogers announced a general standard for causation in FELA cases not one applicable exclusively to injuries involving multiple causes. Id. at 2639. The “in part” language applied as well to the “directness or foreseeability of the connection between the carrier's negligence and the plaintiff's injury.” Id. Therefore, the district court erred in stating that the import of the “in part” standard was merely to hold the carrier liable in cases of negligence by multiple actors. Rogers made clear that the common law consideration about whether a particular cause was “sufficiently substantial” to constitute a proximate cause was replaced with the straightforward “any part” language as the “ single ” inquiry determining causation in FELA cases. Id. at 2638–39. Accordingly, the FELA “in part” standard impacts the causation analysis beyond cases in which the employee is also negligent. That said, the district court properly noted that FELA does not render employers strictly liable for any workplace injury without any showing of negligence. The relevant inquiry, then, is whether the evidence here raised a genuine issue of fact that Metra's negligent breach of duty was a cause, even in the slightest, of the injury to Lynch.

B.

The district court assumed that Lynch had raised a genuine issue of fact as to Metra's breach of its duties under FELA to provide employees a reasonably safe workplace, safe equipment, proper training, and suitable methods to perform the assigned work. Lynch alleged that those duties were breached in three aspects: (1) Metra failed to provide adequate training in fence installation; (2) Metra failed to adopt and enforce reasonably safe work methods and procedures; and (3) Metra failed to inspect, discover and remedy unsafe conditions. Lynch's theory was that the top rail slipped from its cup either because it was cut too short or not securely tightened, or because it was not installed in a manner that appropriately accounted for the steep grade of the hill.

In the district court, Lynch presented evidence that the foreman at the worksite had an ongoing obligation to inspect the work being performed at each phase, and that in the foreman's absence the assistant foreman had that responsibility. At the time of the incident, the foreman was away from the jobsite acquiring needed materials, but the assistant foreman, Reyes, was present. There was also testimony that crew members individually had an obligation to inspect.

In addition, the record contained testimony from some of the crew members as to the procedures for measuring and cutting the top rail so that it fit snugly, as deeply-seated into the cups attached to the fence posts as possible. Scott had worked for 18 or 19 years as a B & B mechanic at Metra and installed fences regularly during that time. He testified that a rail that is cut to the proper length and tightened in the cup should not be able to come out of the cup. Otero, who had 11 years of experience working for Metra as a B & B mechanic, similarly testified that a rail that is cut to the proper length and secured in the cup should not be able to fall out. He further opined that the grade of the hill might have impacted it. He testified that in installing the mesh on the posts and rails, there was concern regarding the impact of the grade of the hill, and that a trench was dug in front of one of the posts in the area of the accident in order to accommodate for the impact of the grade of the hill on the ability to properly secure the mesh to the posts and rail. Finally, Lynch presented evidence that although workers received training on a number of mechanic tasks, they received no formal training for installing fences, although they engaged in that task on a regular basis. They learned how to install fences “on the job,” from other crew members involved in the installation. Although Otero had been a mechanic for 11 years, he had never worked on a fence with an elevation difference like the one involved here, which he described as at least a 2–3 foot drop between posts. He received no training on installing fencing in circumstances such as that one.

That evidence is sufficient to present a genuine issue of fact concerning the causation issue. From that testimony, a jury could reasonably conclude that the top rail fell out because it was either cut too short or improperly tightened in the cup by a Metra employee. The jury could further conclude that the problem would have been discovered if a Metra employee had inspected the top rail after it was installed and before the next phase, fastening the mesh to the skeleton, was initiated. Finally, a jury could determine that the failure to provide training in fence installation left the crew members ill-equipped to adjust to non-standard conditions such as the steep grade of the hill, and that the inability of the employees to anticipate the impact of that grade on the rail contributed to the fall.

C.

The district court held that Lynch could not proceed because those theories were based on nothing more than rank speculation. The court rejected the testimony of Otero as unhelpful because he testified both that a top rail should not come out if cut to the proper length and that the grade of the hill could have caused it to fall out. The court then declared that “more importantly,” Otero was a fact witness not an expert. The court criticized Lynch for failing to measure the distance between the posts, the length of the top rail, and the grade of the hill, and for failing to attempt to recreate the accident “or otherwise investigate or exclude other possible causes.” Dist. Ct. Op. at 7. As to Otero's testimony that the hill was steep and at least a 15 percent grade, the court dismissed it as explicitly speculative.

The district court erred in dismissing the testimony as speculative and demanding direct evidence of the cause of the fall and exclusion of other possible causes. In its statement of uncontested facts for the summary judgment motion, Metra admitted that the crew members were working on a hill, that it was a warm, clear and sunny day and lighting was not a problem, that Lynch was installing the mesh and nothing he did caused the pole to fall, and that no one slipped or fell causing the top rail to dislodge. In other words, Metra acknowledged that the top rail did not fall out as a result of inclement weather or an “act of God,” that Lynch did not cause it to dislodge in his actions installing the mesh, and that the rail did not have any apparent design or manufacturing defect. That leaves the most obvious cause of the fall—the failure to cut it long enough to ensure that it remained seated in the cups, or the failure to securely tighten the cup. Either of those conditions could be easily ascertained if the rail had been inspected prior to proceeding with the next phase of fence construction—installing the mesh fabric. The testimony indicated that a properly installed top rail should be snug, and could move in the cups only slightly; a person inspecting the top rail could have checked the amount of movement and the amount of resistance in its movement, thus determining whether it was properly seated in the cups.

The district court in effect held that the jury could not draw the most obvious conclusion as to the cause of the injury, because there is no direct evidence of that cause and no expert testimony supporting that conclusion. That is inconsistent with the consistent holdings of this and other courts that under FELA, circumstantial evidence alone can support a jury verdict, and expert testimony is unnecessary where the matter is within the realm of lay understanding and common knowledge.

D.

We consider first the implication that expert testimony is necessary to survive summary judgment on a FELA claim. Courts have consistently rejected that position, holding that expert testimony is not required. For instance, in Harbin, we considered Harbin's FELA action against Burlington Northern Railroad, claiming that the unsafe work conditions caused his heart attack. Harbin, 921 F.2d 129. The evidence demonstrated that the roundhouse in which Harbin worked had no special ventilation system. Id. at 129–30. Locomotives left running in the building created clouds of exhaust fumes. Id. at 130. In addition, once a year Harbin would clean the inside of the boilers, by scraping the soot off the inside and blowing it out with an air pressure hose. Id. at 129–30. That would send additional soot into the air. Id. The railroad provided Harbin with a mask but it covered only his mouth and not his nose. Id. at 130. Harbin provided expert testimony from a doctor that inhalation of particulate matter could irritate the lungs and stress the heart, precipitating a heart attack, but did not provide any expert testimony as to the air quality or the amount of soot in the roundhouse air. Id. The railroad protested that Harbin's evidence thus amounted to nothing more than pure fantasy, containing“less substance than broth brewed from the bones of a starved pigeon.” Id. at 131.

The district court in Harbin held there was enough evidence of causation to go to a jury given the medical expert testimony, but that there was insufficient evidence of negligence. Id. The court opined that without knowledge of the precise quantity or composition of soot in the air, a jury would not be able to assess the reasonableness of the railroad's conduct. Id.

Although recognizing that expert testimony would undoubtedly enhance Harbin's case, we held on appeal that it was not essential under the regime of FELA. Id. at 131. We held that “[a] long line of FELA cases reiterate the lesson that the statute vests the jury with broad discretion to engage in common sense inferences regarding issues of causation and fault.” 921 F.2d at 132, citing Rogers, 352 U.S. at 510, 77 S.Ct. 443 (noting that the decision must be left for the jury “in all but the infrequent cases where fairminded jurors cannot honestly differ whether fault of the employer played any part in the employee's injury”). Accordingly, we held that a jury could reasonably conclude that the failure to implement a different cleaning method such as a vacuum rather than air pressure hose, and the failure to take other precautions such as more effective face masks, was negligent. Id. at 131–32. We did not require expert testimony regarding the efficacy or practicality of such measures in order to allow the case to proceed to the jury, noting that “numerous FELA actions have been submitted to a jury based upon far more tenuous proof—evidence scarcely more substantial than pigeon bone broth.” Id. at 132.

Similarly, in Ulfik v. Metro–North Commuter R.R., 77 F.3d 54, 59 (2d Cir.1996), the court considered whether expert testimony was necessary to establish whether exposure to paint fumes on July 15 caused Ulfik's dizziness eight days later on July 23, which was the only disputed link in the causal chain. The court held that a jury could properly infer that exposure to paint fumes caused headaches, nausea and dizziness without the need for expert testimony. Id. at 59–60. The court noted that expert testimony may be necessary where some special expertise is necessary to draw a causal inference because of its esoteric nature, but that in general the causal sequence can be inferred from circumstantial evidence, expert testimony, or common knowledge. Id. at 60, citing W. Page Keeton et al., Prosser & Keeton on the Law of Torts, § 41, at 270 (5th ed. 1984); see also Myers v. Illinois Central R.R. Co., 629 F.3d 639, 643 (7th Cir.2010) (expert testimony unnecessary in cases where the layperson can understand what caused the injury); Wills v. Amerada Hess Corp., 379 F.3d 32, 46 (2d Cir.2004) (in FELA action, expert testimony necessary only if causal link is beyond the knowledge of the lay juror, such as the link between exposures to toxins and squamous cell carcinoma).

The district court thus erred in dismissing Otero's testimony as merely a fact witness not an expert. There was no reason for expert testimony on the easily understood causal connection between improper installation of a top rail and its subsequent drop to the ground. In fact, the inference is an easier one to make than the inferences in Ulfik that exposure to paint fumes caused dizziness eight days later, or the inference in Harbin that the soot stirred up by the idling locomotives and boiler cleaning was so significant as to create a safety concern necessitating additional action by the railroad.

E.

The other basis—and perhaps overriding concern—of the district court appearsto be that Lynch has failed to present any direct evidence establishing the cause of the top rail collapse. Courts have repeatedly held, however, that in FELA cases the element of causation may be established through circumstantial evidence or common knowledge, and that direct or expert testimony is not required. Missouri Pacific R.R. Co. v. Kansas Gas and Elec. Co., 862 F.2d 796, 800 (10th Cir.1988) (a case can rest entirely on circumstantial evidence and still be sufficient to reach the jury under FELA); Gibson v. Elgin, Joliet & Eastern Ry. Co., 246 F.2d 834, 837 (7th Cir.1957) (burden met if proof, though entirely circumstantial, from which a jury may with reason make the inference). In fact, in Rogers the Supreme Court declared that “[c]ircumstantial evidence is not only sufficient, but may also be more certain, satisfying and persuasive than direct evidence.” Rogers, 352 U.S. at 508, n. 17, 77 S.Ct. 443.

The Supreme Court's decision in Gallick v. Baltimore and Ohio R. Co., 372 U.S. 108, 83 S.Ct. 659, 9 L.Ed.2d 618 (1963), which was recently reaffirmed in McBride, is instructive. See McBride, 131 S.Ct. at 2639. Gallick was a spotting crew foreman working on the railroad's right of way when he was bitten by an insect. Gallick, 372 U.S. at 109, 83 S.Ct. 659. In an unfortunate progression, the wound from the bite became infected, and the infection spread throughout his body, resulting in the eventual amputation of both of his legs. Id. None of the doctors who treated Gallick could explain the etiology of his condition, but some of them characterized it as secondary to an insect bite. Id. at 109–10, 83 S.Ct. 659. Gallick filed suit against the railroad under FELA, claiming that the insect bite occurred as he was working near a fetid pool containing dead and decaying rats and pigeons, which had existed for many years and of which the railroad had knowledge. Id. at 100, 83 S.Ct. 659. He argued that the pool of stagnant water attracted insects and resulted in the bite and subsequent infection. Id. The appellate court in the case held that a jury could not reasonably find liability, because there was no direct evidence that the insect had any connection with the pool of water or evidence which would negate alternative possibilities that the insect “had emanated from ‘the nearby putrid mouth of the Cuyahoga River, or from the weeds, or unsanitary places situated on property not owned or controlled by the railroad.’ ” Id. at 112, 83 S.Ct. 659. The appellate court concluded that the evidence was merely a series of guesses and speculations—a chain of causation too tenuous to support liability. Id.

The Supreme Court reversed that determination, holding that the appellate court improperly invaded the function and province of the jury, and that there was sufficient evidence to warrant the jury's conclusion that the injuries were caused by the railroad's acts or omissions. Id. at 113, 83 S.Ct. 659. Specifically, the Supreme Court held that the appellate court erred in requiring either direct evidence that the insect had a connection to the fetid pool, or more substantial circumstantial evidence than that the pool created conditions that furnished an environment to attract and infect such insects. The Court noted that in FELA cases, the role of the court is not to search the record for conflicting circumstantial evidence and to take the case from the jury because the evidence equally supports inconsistent and uncertain inferences. Instead, it is the function of the jury, not the court, to select among conflicting inferences and conclusions.

The Court reached a similar conclusion in Lavender v. Kurn, 327 U.S. 645, 646, 66 S.Ct. 740, 90 L.Ed. 916 (1946), a FELA case alleging that a switchtender's death was attributable to railroad negligence. The switchtender, Haney, was found unconsciousnear the track and died as a result of a fractured skull. Id. at 648, 66 S.Ct. 740. An autopsy revealed an injury to the back of his head made by a fast moving small round object. Id. at 648–49, 66 S.Ct. 740. The petitioner's theory was that Haney was struck by the end of a mail hook hanging down loosely on the outside of a mail car on a backing train. Id. at 649, 66 S.Ct. 740. The petitioner introduced evidence that the mail hook could have swung out 12 to 14 inches, and if it so extended and if Haney was standing on top of a nearby mound of dirt, he could have been struck by the mail hook. Id. The respondent countered that the mound was 10 to 15 feet north of the track and therefore the hook could not have reached Haney's head. Id. at 649–50, 66 S.Ct. 740. Instead, the respondent theorized that Haney was murdered, and introduced evidence that hoboes and tramps frequented the area at night and that Haney carried a pistol to protect himself. Id. at 650, 66 S.Ct. 740. The Court held that there was sufficient evidence of negligence to justify submitting the case to the jury. Id. at 652, 66 S.Ct. 740. The Court held that there was evidence from which it might be inferred that the mail hook struck Haney. Id. The Court acknowledged that there was also evidence indicating that it was physically and mathematically impossible for the hook to strike Haney, and that there were facts from which one could infer that he had been murdered. Id. But the evidence indicating the hook could have reached Haney was sufficient to allow the case to go to the jury. Id. The Court explicitly rejected the notion that the speculative nature of the inquiry should prevent submission of the case to the jury:

It is no answer to say that the jury's verdict involved speculation and conjecture. Whenever facts are in dispute or the evidence is such that fair-minded men may draw different inferences, a measure of speculation and conjecture is required on the part of those whose duty it is to settle the dispute by choosing what seems to them to be the most reasonable inference. Only when there is a complete absence of probative facts to support the conclusion does a reversible error appear.
Id. at 653, 66 S.Ct. 740.

Those cases establish that circumstantial evidence is sufficient to establish FELA liability, and that a jury can make reasonable inferences based on that circumstantial evidence even where conflicting inferences are also appropriate and no direct evidence establishes which inference is correct. See also DeBiasio v. Illinois Central R.R., 52 F.3d 678, 684 (7th Cir.1995) (testimony by worker that, based on his 13 years of experience, the sequence of events indicated that the cars had made impact but failed to couple automatically, enough to submit case to the jury even though no one witnessed the actual impact). Mendoza v. Southern Pacific Transportation Co., 733 F.2d 631, 633 (9th Cir.1984) (slight evidence is sufficient in FELA cases to raise a jury question, and it is only necessary that the conclusion be one that is not outside the possibility of reason on the facts and circumstances shown); Gibson, 246 F.2d at 837 (jury verdicts can be based solely on speculation, conjecture and possibilities).

F.

Finally, the district court and Metra rely on Coffey v. Northeast Illinois Regional Commuter R.R. Corp. (METRA), 479 F.3d 472 (7th Cir.2007) for the assertion that conjecture is insufficient to avoid summary judgment, but that is demonstrably not the holding in Coffey. In fact, Coffey further reinforces Lynch's argument that summaryjudgment was improperly granted. In Coffey, the plaintiff, a train engineer, asserted a FELA claim based on an injury sustained when his head struck a sun visor as he climbed into the driver's cab of the train. Id. at 474. Coffey hypothesized that the bolts fastening the visor to the wall had been loosened by the vibration of the train, causing the visor to descend halfway so that it was pointing at the driver's head. Id. at 475. Rather than dismiss that possibility as speculative, we held that “[t]he conjecture is implausible, though not quite so outlandish that it can be rejected as a matter of law. But pretty outlandish....” Id. In finding sufficient evidence of causation, we noted that “it is possible to tell a story in which the horizontal position of the visor in this case was the result of the railroad's negligence in failing to tighten the bolts.” Id. at 476. Recognizing that evidence may be merely circumstantial, we opined that it might even be argued that the position of the visor was itself evidence of negligence, as the position has no utility. Id. at 477. Therefore, in Coffey we did not fault the plaintiff for failing to present sufficient evidence of causation or for relying on circumstantial evidence. Coffey in fact held that the causation theory could not be dismissed as a matter of law even though the possible story was outlandish. The problem in Coffey instead was with the failure to provide any evidence at all of foreseeability. No evidence was presented of the proximity of the visor to the driver's head when groping for the light switch, or of the weight and padding of the visor, either of which would have given the railroad reason to foresee injury and take precautions. Id.

Similar to Coffey, the fall of the pole from the cups holding it is itself evidence that the pole was not properly installed—either as a result of a failure to properly cut the pole to the optimal length or to secure the pole tightly in the cup. The testimony of Otero and Scott was consistent that a pole cut to the proper length and securely tightened will not fall. Although the district court appeared to believe that Otero contradicted that testimony in also stating that the hill may have caused the fall, there is nothing inherently inconsistent in those two statements. The effect of gravity could certainly cause a pole on the lower post to seat itself more deeply in the lower cup, with the result that a pole that was cut even slightly too short or not tightened securely would fall. There is nothing inconsistent in identifying both the hill and the improper installation as causing the pole to fall. Moreover, the failure to inspect the pole installation before workers proceeded to install the mesh—an inspection that Metra appears to require—establishes causation as well because there is no reason to believe that an improperly installed pole would not have been discovered in the course of an inspection. The failure to inspect the installation of the top rail was particularly problematic because the testimony indicated that the work crew was concerned about the installation of the fence given the steep grade of the hill, and that the top rail for that section was left uncompleted because of the difficulties it presented. Those difficulties arguably were compounded by the failure to train the employees as to how to address such circumstances. Thus, it is not only “possible to tell a story” that involves employer negligence here, it is in fact the most likely explanation for the events given the uncontradicted testimony that a properly cut and tightened pole will not fall, and that workers were supposed to inspect the work at each stage. And the concern in Coffey with the lack of evidence of foreseeability is not even an issue raised by the parties in this case—nor should it be. The danger of an improperly secured metal pole suspended in the air above employees working to secure a mesh fence is obvious, and the need to take precautions follows from that. In fact, the obligation to inspect the work at each stage is undoubtedly in recognition of that danger.

Lynch therefore adequately raised genuine issues of material fact as to negligence as required under FELA. The decision of the district court granting summary judgment is Vacated and the case Remanded for further proceedings consistent with this opinion.