BYE, Circuit Judge.

Carol O'Neal sued Remington Arms Company, L.L.C. (Remington) alleging a defect in a rifle manufactured by Remington caused her husband's death in a November 2008 hunting accident. The district court granted summary judgment to Remington on the grounds that O'Neal could not show whether the alleged defect existed at the time of manufacture or whether the defect resulted from a subsequent alteration or modification to the rifle. Because South Dakota law permits a plaintiff to prove a product defect through circumstantial evidence, and O'Neal presented sufficient circumstantial evidence to show the alleged defect was present at the time of manufacture and was not the result of a subsequent alteration or modification, we reverse and remand for further proceedings.

I

In this appeal from the grant of a summary judgment, we recite the facts in the record in the light most favorable to O'Neal, giving her the benefit of all reasonable inferences to be drawn from the evidence. See Turner v. Iowa Fire Equip. Co., 229 F.3d 1202, 1204 (8th Cir.2000).1

Remington manufactured the Model 700.243 caliber bolt-action rifle involved in this case in 1971. Remington used a trigger mechanism in its Model 700 rifles called the Walker trigger, named after the engineer who designed the mechanism. Remington knows the Walker trigger can cause Model 700 rifles to fire a round when the safety lever is released from the safe position to the fire position, without the trigger being pulled. This defect results from the manner in which two components of the trigger mechanism — the sear and the connector — interact with one another, coupled with the lack of a physical attachment between the connector and the trigger itself.

We start with an explanation of the interaction between the sear and the connector. The connector is a U-shaped piece of steel found directly in front of the trigger. The connector gets pushed forward when the trigger is pulled. The sear is a separate metal piece which rests on the very tip of the rear corner of the connector. When the connector moves forward, the sear drops down behind it. The motion of the sear dropping allows the firing pin to snap forward, and the rifle fires a cartridge. Even when the connector is properly aligned, the amount of engagement (or overlap) between the rear corner of the connector and the sear is very small, just.01 to .025 inches. In other words, the inherent design of the Walker trigger allows a Model 700 rifle to fire a cartridge when the connector is pushed forward as little as 1/100th to 25/1000ths of an inch, allowing the sear to drop.2 Figures 1 and 2 in the attached appendix show the relationship between the trigger, sear, and connector, as well as the minute engagement point between the connector and the sear.

Next, we explain how the lack of a physical connection between the trigger and the connector makes the minute engagement point between the connector and the sear so critical, and susceptible to malfunction. In the Walker trigger mechanism, the connector and the trigger are not physically attached; rather, the connector is "slip fit" in front of the trigger piece. As a result, the two parts separate slightly every time a rifle is fired. Very small pieces of dirt, manufacturing residue, corrosion deposits, lubricant deposits, firing deposits, and even condensation can get trapped between the connector and the trigger when the two parts separate. The connector is not properly aligned when this happens, because the foreign material in the small space between the two parts pushes the connector forward.

When the safety lever is in the safe position, it physically lifts and restrains the sear out of engagement with the connector and trigger. But when the safety lever is in the fire position, the minute engagement point between the rear corner of the connector and the sear becomes critical, because that small engagement point is the only thing keeping the sear from dropping and allowing the firing pin to snap forward. If the connector is misaligned by as little as 1/100th of an inch, the tip of the connector's rear corner no longer supports the sear above it. As a result, the sear can drop behind the connector without the trigger being pulled, with only the safety lever lifting and restraining the sear. Thus, when someone releases the safety lever, the rifle fires a round without the trigger being pulled. See Support Services Engineering Report prepared by Charles W. Powell (Powell Report), Jt. App. at 579-86.

Remington knew about this problem with the Walker trigger at least as early as 1979. The record in this case includes the minutes from a Remington product safety subcommittee meeting dated January 2, 1979. Jt.App. at 624-28. The subcommittee minutes discuss the inspection procedures Remington initiated on all bolt action rifles beginning in 1975, including the Model 700. Based in part upon the inspection of rifles returned to Remington for repairs, Remington acknowledged that Model 700 rifles manufactured prior to 1975 can be "`tricked' into firing when the safety lever is released from the `safe' position" without pulling the trigger. Id. at 624. Remington estimated that at least 1% of the two million Model 700 rifles it had manufactured prior to 1975 — or 20,000 rifles — would inadvertently fire merely by releasing the safety (i.e., moving the lever from the safe position to the fire position) without pulling the trigger. Id. at 627. Remington decided against recalling the Model 700 rifles, though, because "the recall would have to gather 2,000,000 guns just to find 20,000 that are susceptible to this condition." Id. at 627.

When viewed in the light most favorable to O'Neal, however, the record suggests the lack of a physical attachment between the trigger and connector in the Walker trigger creates the possibility of foreign material getting trapped in the space between the two parts every time a Model 700 rifle is fired. If foreign material is present and pushes the connector far enough forward — past its already minute engagement point with the sear — any Model 700 rifle could be susceptible to an inadvertent discharge at some point. Moreover, the sear and connector in Model 700 rifles are enclosed in a riveted housing that interferes with a user's ability to visually inspect the interior parts to determine whether the connector has an insufficient engagement with the sear due to the presence of foreign materials trapped between the trigger and connector. See Powell Report, Jt.App. at 584.

On November 9, 2008, O'Neal's husband, Lanny, was deer hunting with friends near Eagle Butte, South Dakota. Lanny loaned the Remington Model 700.243 caliber bolt-action rifle involved in this case to Mark Ritter, another one of the hunters, to use that day. The hunters were traveling in a pickup truck when they spotted a deer. At the time, Ritter was sitting in the back seat of the pickup behind Lanny, who sat in the front passenger seat. After the pickup stopped, Ritter began to exit the truck to shoot the deer. Ritter moved the safety lever on the rifle from the safe position to the fire position without pulling the trigger, and the rifle discharged. The cartridge traveled through the pickup seat and hit Lanny, who eventually died from the gunshot.

As stated above, the particular Model 700 rifle involved in this case was manufactured in 1971. Remington does not dispute that, at the time of its manufacture, the rifle was equipped with a Walker trigger mechanism. Remington does dispute, however, whether O'Neal can prove the rifle never underwent a post-manufacture alteration or modification which could have caused it to discharge when the safety lever was moved from the safe position to the fire position without pulling the trigger. The evidence marshaled by O'Neal shows that Doug Swanson, Lanny's stepfather, acquired the rifle in the mid-1980s from the estate of a man named Albert McIlvenna. Swanson thereafter used the rifle and occasionally loaned it to others, including Lanny. Lanny borrowed the rifle from Swanson in 2005 or 2006, and remained in possession of it until the day of the 2008 accident. Swanson never altered or modified the rifle during the time he owned the rifle and has no knowledge of anyone else having altered it. Significantly, no one who used the rifle following Swanson's acquisition of it ever experienced an inadvertent discharge when the safety lever was moved from the safe position to the fire position without pulling the trigger.

Thus, the evidence viewed in the light most favorable to O'Neal indicates the rifle did not undergo any alterations or modifications which would cause it to discharge when the safety lever was moved from the safe position to the fire position without pulling the trigger, at least from the time Swanson first acquired it until the day of the accident. But this leaves a gap of over a decade of unaccounted time between the date of the rifle's manufacture and Swanson's acquisition of it. This gap in time is the center of the parties' current dispute. Remington argues, and the district court agreed, that O'Neal cannot show whether a modification or alteration to the rifle prior to Swanson's acquisition of it could have caused it to discharge when the safety lever was moved from the safe position to the fire position without pulling the trigger.

O'Neal's burden of showing the rifle never underwent a post-manufacture alteration or modification is complicated by the fact that the rifle can no longer be examined or tested. After her husband's death, O'Neal contacted two attorneys to explore the possibility of pursuing a wrongful death action. When both attorneys declined to represent her, and she did not believe a suit against Remington was a viable possibility, she asked a friend to destroy the rifle because it reminded her of her husband's death. Several months after destroying the rifle, O'Neal learned of the defects in the Walker trigger used in Remington's Model 700 rifles after hearing about a national news documentary which aired on CNBC in October 2010. The news documentary includes a videotape prepared by police snipers in Portland, Maine, showing a Model 700 rifle discharging several rounds when an officer merely touches the bolt to unload the weapon when the safety is in the fire position, without touching the trigger. The documentary also features a number of inadvertent discharges of Model 700 rifles which resulted in deaths or serious injuries when the safety lever was switched from the safe position to the fire position without the trigger being pulled. See CNBC Remington Under Fire DVD, Exhibit T, Jt.App. at 1159. O'Neal then contacted a different attorney and shortly thereafter filed this action against Remington in December 2011.

The district court granted summary judgment to Remington on the grounds that O'Neal could not show the defect that allegedly caused her husband's death was present at the time of manufacture, and not the result of a post-manufacture alteration or modification which would cause the rifle to discharge when the safety lever was moved from the safe position to the fire position without the trigger being pulled. O'Neal then filed this timely appeal.

II

We review the district court's grant of summary judgment de novo. Occidental Fire & Cas. Co. v. Soczynski, 765 F.3d 931, 935 (8th Cir.2014).

This diversity case is governed by South Dakota law. South Dakota law allows a plaintiff in a products liability suit to use circumstantial evidence to prove that a defective product caused an injury, and that the defect existed when the product left the defendant's control. See Crandell v. Larkin and Jones Appliance Co., Inc., 334 N.W.2d 31, 34 (S.D.1983) ("Causation may be established by circumstantial evidence where that evidence establishes by a preponderance, the probability that the accident was caused by a defect. We do not require that plaintiff eliminate all other possible explanations of causation that the ingenuity of counsel might suggest. It is sufficient that plaintiff negate his own and others' misuse of the product. In addition plaintiff has the burden of showing that the defect existed when the product left [defendant's] hands. This burden may also be satisfied by circumstantial evidence." (quoting Shaffer v. Honeywell, Inc., 249 N.W.2d 251, 256 (S.D. 1976))).

As part of the burden of showing the defect existed when the product left the defendant's control, a plaintiff must show the product reached the plaintiff without any alterations or modifications "which defeat the safety which is engineered into that product." Peterson v. Safway Steel Scaffolds Co., 400 N.W.2d 909, 914 (S.D.1987); see also S.D. Codified Laws § 20-9-10 ("No manufacturer ... of a product may be held liable for damages for personal injury ... where a proximate cause of the injury ... was an alteration or modification of such product ...").

O'Neal argues the circumstantial evidence in this record is sufficient to show the Model 700 rifle involved in her husband's death was (1) defective at the time it left Remington's control, and (2) did not undergo any post-manufacture alterations or modifications which would cause it to discharge when the safety lever was moved from the safe position to the fire position without the trigger being pulled.

With respect to the first point — whether the Model 700 rifle was defective at the time it left Remington's control — Remington does not dispute the Model 700 rifles it manufactured in 1971 were equipped with the Walker trigger. The record in this case establishes that the problems with the inherent design of the Walker trigger are well-documented and well-known to Remington. Indeed, Remington itself acknowledged that at least 20,000 rifles it manufactured prior to 1975 were susceptible to inadvertent discharges when the safety lever was moved from the safe position to the fire position without the trigger being pulled.

A reasonable inference of the additional evidence presented by O'Neal through the plenary report and testimony of her expert, Charles Powell, is that any Model 700 rifle equipped with a Walker trigger may be susceptible to an inadvertent discharge when the safety lever is moved from the safe position to the fire position without the trigger being pulled. The Walker trigger, as designed, allows the connector and trigger to separate when the rifle is fired, creating the possibility of foreign material getting trapped between the trigger and connector, which misaligns the connector by pushing it forward. This, coupled with the already minute engagement point between the sear and the connector, can result in a Model 700 rifle discharging without the trigger being pulled when the connector is misaligned by as little as 1/100th of an inch, or the thickness of 2½ pieces of standard copy paper.

Moreover, the Walker trigger hides this latent defect inside a riveted housing unit which interferes with a user's ability to clean the interior parts to remove the presence of foreign materials, or to visually inspect the parts to determine whether the connector has become misaligned and has an insufficient engagement with the sear. All of this makes it very difficult to prevent an inadvertent discharge from occurring in a Model 700 rifle; more significantly, it makes it very difficult to predict when an inadvertent discharge caused by this design defect may occur. We agree with O'Neal that a rifle originally manufactured in this condition, which allows for the possibility of the rifle discharging without pulling the trigger, is defective and not fit for its ordinary purpose. See Powell Report, Jt.App. at 586; cf. Lewy v. Remington Arms Co., 836 F.2d 1104, 1106-08 (8th Cir.1988) (reviewing a products liability action against Remington which included a claim alleging this same defect in Model 700 rifles and noting "there was sufficient evidence from which the jury could find that Remington knew the [Model 700 rifle] was dangerous").

The second point — whether the Model 700 rifle involved in this case underwent a post-manufacture alteration or modification which would cause it to discharge when the safety lever was moved from the safe position to the fire position without the trigger being pulled — is the more difficult issue. O'Neal argues the evidence from the men who owned and used the rifle without incident from the mid-1980s through the date of the accident is sufficient circumstantial evidence to show the rifle did not undergo a post-manufacture modification or alteration prior to the mid-1980s. In addition, O'Neal relies upon the opinion of her expert, Charles Powell, who testified that post-manufacture modifications or alterations to the rifle, such as an improper adjustment of the sear engagement, would most likely cause the rifle to fire every time the safety was released without a trigger pull. See Powell Deposition at 55-56; Jt.App. at 1121. We agree this is sufficient circumstantial evidence under South Dakota law to survive summary judgment.

For purposes of summary judgment, we accept as true O'Neal's evidence that the rifle inadvertently discharged in November 2008 when Ritter moved the safety lever from the safe position to the fire position without pulling the trigger. Starting from this premise, the circumstantial evidence marshaled by O'Neal supports a reasonable inference that the rifle did not undergo any post-manufacture alterations or modifications which would cause it to discharge when the safety lever was moved from the safe position to the fire position without the trigger being pulled. The fact that the subject rifle was used many times without incident from the mid-1980s through November 2008, and then suddenly inadvertently discharged, is consistent with the unpredictable manifestation of the inherent design defect in the Walker trigger. In sharp contrast, if the subject rifle had been modified or altered prior to the mid-1980s in a way which would cause it to discharge when the safety lever was moved from the safe position to the fire position without the trigger being pulled, it is highly unlikely the rifle could have been used as many times as it was over the span of the next twenty-plus years without incident. As O'Neal's expert testified, an alteration or modification prior to Swanson's acquisition of the rifle which would cause it to discharge when the safety lever was moved from the safe position to the fire position without the trigger being pulled would have caused the rifle to inadvertently discharge "probably almost every time" it was used. Jt.App. at 1121. That fact is clearly inconsistent with the evidence O'Neal presented of the rifle's usage history.

Remington argues O'Neal still cannot show the subject rifle had the same Walker trigger at the time of the accident that it had when it left Remington's control in 1971. Remington notes "that original factory trigger mechanisms are often replaced with trigger mechanisms designed, manufactured and sold by a variety of after-market trigger manufacturers." Appellee's Br. at 25. Remington argues the destruction of the rifle precludes O'Neal from establishing that the trigger mechanism in the subject rifle at the time of the accident was the original Remington trigger mechanism rather than an after-market trigger made by another manufacturer. We disagree.

First of all, the record shows the hunting accident was investigated by the Dewey County Sheriff's Office; the South Dakota Department of Game, Fish and Parks; and the Cheyenne River Sioux Tribe. These investigations included inspections of the rifle, including at least one description of the various parts of the rifle which specifically mentions additions to the rifle made by other manufacturers. See Jt.App. at 240 (noting the presence of a Simmons scope). Thereafter, the rifle was transferred to the Federal Bureau of Investigation (FBI). None of the records from the law enforcement agencies who inspected the rifle or had possession of it following the accident reference the presence of an after-market trigger mechanism. The absence of such information supports the reasonable inference that the trigger mechanism in the subject rifle was the original Remington trigger. Cf. Powell Deposition at 64; Jt.App. at 1123 ("[L]aw enforcement investigators looked at the rifle, cycled the bolt, took the shells out of it. Had there been any broken parts in the fire control or any other part of the rifle you would have thought that would have been noted, by the absence of that information, you would think there were no broken parts.").

Second, and perhaps more importantly, the after-market triggers referenced by Remington are designed to eliminate the possibility that a Model 700 rifle will inadvertently discharge when the safety lever is moved from the safe position to the fire position without the trigger being pulled, not cause the condition to occur. Here, there was no evidence the subject rifle had a history of inadvertent discharges that might spur an owner to replace the trigger mechanism. In addition, as stated above, for summary judgment purposes we accept as true O'Neal's evidence that the rifle inadvertently discharged in November 2008 when Ritter moved the safety lever from the safe position to the fire position without pulling the trigger. Starting from this premise, the fact that the subject rifle was used many times without incident from the mid-1980s through November 2008, and then suddenly inadvertently discharged, is more consistent with the unpredictable manifestation of the inherent design defect in the Walker trigger, than it is with the rifle being equipped with a replacement trigger designed to eliminate the possibility of an inadvertent discharge.

While a jury may ultimately find this evidence lacking, for our purposes in reviewing the grant of summary judgment, we agree with O'Neal that the reasonable inferences supported by this record require us to reverse and remand for further proceedings. Cf. Ahmann v. United Air Lines, Inc., 313 F.2d 274, 281 (8th Cir. 1963) (reviewing the improper grant of a motion for judgment notwithstanding the verdict and noting "[i]t is not the function of a court to search the record for conflicting circumstantial evidence in order to take the case away from the jury on a theory that the proof gives equal support to inconsistent and uncertain inferences.... The very essence of [the jury's] function is to select from among conflicting inferences and conclusions that which it considers most reasonable.") (citation omitted).

Finally, Remington urges us to affirm the district court on two alternative grounds: (1) that Powell's expert testimony should be excluded under Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993); and (2) that the case should be dismissed as a sanction for spoliation of evidence because O'Neal destroyed the rifle. The district court did not address either of those issues in its motion for summary judgment. We believe it is more prudent for the district court to address those issues in the first instance, and thus decline Remington's invitation to affirm on either of these alternative grounds. See, e.g., Red River Freethinkers v. City of Fargo, 679 F.3d 1015, 1028 (8th Cir.2012) (declining to reach the merits of the claims where the district court did not rule on the merits and neither party addressed the merits in its brief); Dodd v. United States, 614 F.3d 512, 518 (8th Cir.2010) (remanding to the district court to address the merits of the claims where neither party briefed the merits to the court and the district court did not address the merits in the first instance); Alliant Techsystems, Inc. v. Marks, 465 F.3d 864, 873 (8th Cir. 2006) ("Because the district court did not decide the merits of these claims, which are heavily fact-based, we decline to consider them in the first instance.").

III

We reverse and remand to the district court for further proceedings consistent with this opinion.

LOKEN, Circuit Judge, dissenting.

I respectfully dissent. In my view, the district court correctly concluded that, because Carol O'Neal had the Remington Model 700 rifle destroyed before this action was filed, she cannot prove that a design or manufacturing defect present when the rifle left Remington's control in 1971 proximately caused the tragic hunting accident that killed her husband 37 years later. To make clear why this is so, I will separately analyze two distinct defect theories articulated by O'Neal's expert, Charles W. Powell, which the court has confusingly intermingled.

First, as the court has explained, Powell opined in his October 2012 SSEC Engineering Report that the triggers in all Model 700 rifles manufactured at that time were defective because the "Walker" fire control system included a connector that separates briefly from the trigger when the rifle is fired, permitting "interferences" such as dirt and deposits to become lodged between the connector and trigger, which can lead to a misfire. This is a design defect theory. Crucial to this theory is evidence that an interference was present that caused a particular Model 700 rifle to misfire causing injury. Lacking this evidence, Powell at his deposition disclaimed any reliance on his design defect theory in this case:

Given this testimony by O'Neal's only expert, the summary judgment record contained no evidence that the alleged design defect caused the misfire that killed Lanny O'Neal. Tellingly, while the first three pages of the fact section of O'Neal's brief to this court explained the alleged connector design defect in detail, her brief opposing summary judgment submitted to the district court did not even mention the connector or this design defect theory. An appellate court should always be skeptical when it is asked to reverse based on a claim or theory not presented to the trial court.

Second, as the court has explained, Powell's Report referenced internal Remington documents estimating that 1% of the Model 700 rifles manufactured before 1975 were defectively manufactured with "inadequate Sear lift" that caused the rifle to "inadvertently fire by movement of the safety lever to the OFF position, without pulling the trigger." O'Neal argued to the district court that summary judgment should be denied because there was sufficient circumstantial evidence that the rifle that killed her husband was one of the 1% that had this manufacturing defect — what the parties refer to as a "trick condition" — that causes the rifle to discharge when the safety lever is released without pulling the trigger. Powell's deposition testimony established that inspecting the rifle was critical to proving this theory:

To reverse the grant of summary judgment, we must conclude that O'Neal presented sufficient evidence that, when the Model 700 rifle left Remington's control, it was in an unreasonably dangerous condition because of a design or manufacturing defect, and that Lanny O'Neal's injury was proximately caused by that defect, rather than by a subsequent alteration of the rifle. See Burley v. Kytec Innovative Sports Equip., Inc., 737 N.W.2d 397, 408-09 (S.D.2007). "Expert testimony is generally necessary to establish elements of negligence and strict liability." Nationwide Mut. Ins. Co. v. Barton Solvents, Inc., 855 N.W.2d 145, 151 (S.D.2014), citing Burley, 737 N.W.2d at 407-11.

Here, expert Powell could not opine that the alleged connector design defect caused the shooting without examining the rifle to determine if any foreign debris or other interference had separated the connector from the trigger body. Nor could Powell opine without examining the rifle whether it was one of the 1% of Model 700 rifles manufactured in the early 1970's with the alleged "inadequate Sear lift" manufacturing defect. In addition, he could not testify without examining the rifle whether subsequent alteration, adjustment, or damage proximately caused it to fire without the user pulling the trigger. I agree that South Dakota law permits plaintiffs to prove elements of a products liability claim by circumstantial evidence in many situations. But given expert Powell's testimony, O'Neal's only circumstantial evidence that the alleged design defect caused the accident, or that the gun left Remington's control with the alleged manufacturing defect, was the fact that Ritter, an excited deer hunter using a borrowed rifle, told accident investigators the rifle fired on release of the safety without a trigger pull.

Under South Dakota law, "unless it is patently obvious that the accident would not have happened in the absence of a defect, a plaintiff cannot rely merely on the fact that an accident occurred." Nationwide, 855 N.W.2d at 151 (quotation omitted). Here, whether a design or manufacturing defect present when the Model 700 rifle left Remington's control proximately caused a tragic shooting 37 years later was far from "patently obvious." After giving O'Neal ample opportunity to develop and present probative evidence to establish both defect and causation, the district court concluded "there is a complete lack of evidence such that a jury could only speculate and guess as to issues in which O'Neal bears the burden of proof." On this record, summary judgment was properly granted because, as in Burley, O'Neal "has not provided an evidentiary basis tending to show that the injury was caused by a defect rather than the alteration of the product." 737 N.W.2d at 409.

Accordingly, I would affirm the judgment of the district court.