UNPUBLISHED

UNITED STATES COURT OF APPEALS
                FOR THE FOURTH CIRCUIT


JOHN SAMUEL RILEY, III; SANDRA          
MARIE RILEY, Individually and as
Mother and Next Friend of John
Samuel Riley IV and Joshua
Samuel Riley; MAXINE CECILIA
WILLOUGHBY,
               Plaintiffs-Appellants,             No. 99-2305

                 v.
DE’LONGHI CORPORATION, d/b/a
De’Longhi America,
              Defendant-Appellee.
                                        
           Appeal from the United States District Court
            for the District of Maryland, at Greenbelt.
                 Peter J. Messitte, District Judge.
                          (CA-99-44-PJM)

                      Argued: September 28, 2000

                      Decided: October 30, 2000

      Before NIEMEYER and MICHAEL, Circuit Judges, and
   Frederick P. STAMP, Jr., Chief United States District Judge
 for the Northern District of West Virginia, sitting by designation.



Reversed and remanded by unpublished per curiam opinion.


                             COUNSEL

ARGUED: Leonard I. Rosenberg, ROSENBERG & FAYNE, L.L.P.,
Riverdale, Maryland, for Appellants. Thomas Edward Healy, PINO &
2                 RILEY v. DE’LONGHI CORPORATION

ASSOCIATES, L.L.P., White Plains, New York, for Appellee. ON
BRIEF: Scott L. Needleman, ROSENBERG & FAYNE, L.L.P.,
Riverdale, Maryland, for Appellants. Rudolph V. Pino, Jr., PINO &
ASSOCIATES, L.L.P., White Plains, New York; Timothy L. Mullin,
Jr., Edward W. Brady, MILES & STOCKBRIDGE, Baltimore, Mary-
land, for Appellee.



Unpublished opinions are not binding precedent in this circuit. See
Local Rule 36(c).


                             OPINION

PER CURIAM:

   John Riley and Sandra Riley, Sandra Riley on behalf of their two
children, and Maxine Willoughby (together, the "Rileys") sued the
De’Longhi Corporation (De’Longhi), alleging that a defect in a porta-
ble heater manufactured by De’Longhi caused a fire in the Riley
home. The district court entered summary judgment in favor of
De’Longhi. Because we conclude that the Rileys have presented suffi-
cient circumstantial evidence of a product defect to create a genuine
issue of material fact, we reverse and remand for further proceedings.

                                  I.

   On December 14, 1995, John and Sandra Riley, their two children,
and Mrs. Riley’s mother (Maxine Willoughby) were at the Riley
home in New Carrollton, Maryland, when a fire started on the rear
porch. Although the Rileys managed to escape serious physical
injury, their home was severely damaged, many personal belongings
were destroyed, and one of the children (Joshua) has been diagnosed
with post-traumatic stress disorder.

   At the time of the fire a portable oil-filled space heater manufac-
tured by De’Longhi was located on the rear porch of the Riley home.
An extension cord connected the heater to the wall outlet. After the
                   RILEY v. DE’LONGHI CORPORATION                      3

fire was extinguished, investigator James Laws from the Prince
George’s County Fire Department inspected the house. He observed
beading and breaks in the extension cord and therefore concluded that
an electrical overload in the extension cord caused the fire. The
Rileys’ insurance company, the Blue Ridge Insurance Company (Blue
Ridge), hired an expert, Peter Vallas, to investigate the cause and ori-
gin of the fire. Vallas concluded that the fire originated in the heater,
either from an electrical malfunction or a short circuit in the control
panel or wiring. Blue Ridge brought a subrogation action against
De’Longhi, which was eventually settled.

   The Rileys filed a separate (diversity) action against De’Longhi,
alleging negligence, breach of express warranty, breach of implied
warranty of merchantability, strict liability, and violations of the
Maryland Consumer Protection Act and Consumer Products Guaranty
Act. The Rileys hired Trident Engineering Associates (Trident) to
provide an expert opinion as to the cause of the fire. Trident assigned
the case to electrical engineer Kenneth Fennell and fire investigator
George Meyer. Fennell and Meyer wrote a report attributing the fire
to an electrical malfunction within the heater’s control panel and its
attached wiring. De’Longhi also hired an expert, electrical engineer
Leonard Wharton. Based on his observation of beading on the exten-
sion cord, Wharton concluded that the fire was caused by electrical
and thermal failure in the extension cord.

   De’Longhi moved for summary judgment, arguing that the Rileys’
experts, Fennell and Meyer, could not identify a defect in the heater
and that their opinions were based on conjecture. The district court
granted De’Longhi’s motion after concluding that Fennell’s deposi-
tion testimony was too elusive. The Rileys appeal.

                                   II.

   We review a grant of summary judgment de novo. See Goldstein
v. Chestnut Ridge Volunteer Fire Co., 

, 340 (4th Cir.
2000). Summary judgment is appropriate when there is no genuine
issue as to any material fact and the moving party is entitled to judg-
ment as a matter of law. See Fed. R. Civ. P. 56(c). In considering the
propriety of summary judgment, "[t]he evidence of the non-movant[s]
is to be believed, and all justifiable inferences are to be drawn in
4                  RILEY v. DE’LONGHI CORPORATION

[their] favor." Anderson v. Liberty Lobby, Inc., 

, 255
(1986).

   Under Maryland law a plaintiff in a products liability case must
prove: "(1) the existence of a defect, (2) the attribution of the defect
to the seller, and (3) a causal relationship between the defect and the
injury." Virgil v. "Kash n’ Karry" Serv. Corp., 

, 30,

, 656 (Ct. Spec. App. 1984). This standard applies
whether a claim is characterized as strict liability, negligence, or
breach of warranty. See Watson v. Sunbeam Corp., 

,
387 n.3 (D. Md. 1993). In instances when the plaintiff cannot prove
the existence of a defect through direct evidence, he or she is allowed
to introduce circumstantial evidence from which an inference of a
product defect can be drawn. See Harrison v. Bill Cairns Pontiac, 

, 50, 

, 390 (Ct. Spec. App. 1988). This
method of proving a product defect is sometimes called the "indeter-
minate defect theory."

   De’Longhi makes two arguments: (1) the use of circumstantial evi-
dence is inappropriate in this case and (2) even if circumstantial evi-
dence is proper, the Rileys have not forecast a sufficient
circumstantial case to proceed to trial. We disagree. Because the
heater sustained severe fire damage, it is appropriate for the Rileys to
use circumstantial evidence of a product defect. In addition, the
Rileys have proffered sufficient evidence to allow an inference that
a heater defect was the cause of the fire. They have thus created a
genuine issue of material fact on the issue of whether the heater was
defective.

                                  A.

   De’Longhi first argues that the indeterminate defect theory does
not apply in this case because the heater was not completely
destroyed and was in fact available for inspection. De’Longhi is cor-
rect that the indeterminate defect theory applies when an allegedly
defective product is lost or destroyed in the accident. In such a situa-
tion, circumstantial evidence of a product defect is permitted because
direct evidence may not be available. See Restatement (Third) of
Torts: Products Liability § 3 cmt. b (1997).
                   RILEY v. DE’LONGHI CORPORATION                       5

   In this case, even though the heater was available for inspection,
it was substantially damaged as a result of the fire. The fire essentially
consumed the control panel, the electrical switches, and the thermo-
stat assemblies. The heater sustained such severe fire damage that
direct evidence of a product defect may not be available. One of the
Rileys’ experts, Meyer, testified that he could not identify the precise
failure within the heater because of the extensive damage. Therefore,
this is an appropriate case for allowing the use of circumstantial evi-
dence to prove a product defect.

                                   B.

   De’Longhi argues in the alternative that the Rileys did not offer
sufficient circumstantial evidence of a product defect to proceed to
trial. Maryland courts consider five factors in determining whether a
product defect may be inferred from circumstantial evidence:

     (1) expert testimony as to possible causes; (2) the occur-
     rence of the accident a short time after the sale; (3) same
     accidents in similar products; (4) the elimination of other
     causes of the accident; (5) the type of accident that does not
     happen without a defect.

, 549 A.2d at 390 (citations omitted). See
also 

. When the evidence is viewed in the
light most favorable to the Rileys as nonmovants, at least four of the
Harrison factors tip in their favor. That is sufficient.

    Under the first Harrison factor we must consider whether the
Rileys have offered expert testimony as to possible causes of the acci-
dent. Their experts, Fennell and Meyer, authored a report which iden-
tified the heater as the source of the fire and specified that the
problem was an electrical malfunction within the heater’s control
panel. Fennell testified at his deposition that the fire originated in the
heater and that it was probably due to a manufacturing or design
defect. In addition, Fennell identified the possible causes as a thermo-
stat malfunction or a poor electrical connection. During his deposition
Meyer testified that "the heater was responsible for the ignition source
of this particular fire." Meyer also gave his opinion that the fire began
6                  RILEY v. DE’LONGHI CORPORATION

when plastic components in and around the thermostat ignited and fell
into the control panel.

   De’Longhi argues that the Rileys did not satisfy the first Harrison
factor because their electrical engineering expert, Fennell, failed to
identify a precise defect within the heater. Admittedly, Fennell’s testi-
mony is not a model of clarity and certainty. Although in the begin-
ning of his deposition Fennell clearly stated his opinion that the fire
originated within the heater, he later used more qualified language,
stating that "the most probable cause was a failure inside the heater."
De’Longhi points out that Fennell could not identify the particular
connector that might have malfunctioned and that he could not say
"one way or another for sure" whether the excessive heating he identi-
fied in a photograph was due to an incident within the heater. Never-
theless, De’Longhi’s attempt to discredit Fennell’s testimony fails for
two reasons. First, the Harrison analysis is used to establish a circum-
stantial case when there is no direct evidence of a product defect. See

(noting the "significant distinction . . .
between the evidence necessary to ‘directly prove’ a causal design
defect and the lesser evidence required to meet the ‘possible cause’
factor of the Harrison indirect scheme of proof"). In a circumstantial
case the "plaintiff need not explain specifically what constituent part
of the product failed." Restatement (Third) of Torts: Products Liabil-
ity § 3 cmt. c. Fennell’s inability to identify a precise defect is there-
fore not fatal to the Rileys’ effort to establish their case through
indirect proof. Second, the first Harrison factor only requires testi-
mony as to possible causes. Fennell testified that to a reasonable
degree of engineering certainty, he believed the fire originated within
the heater. He also said that, in his opinion, the cause was either a
thermostat malfunction or a poor electrical connection. Fennell’s tes-
timony is sufficient to satisfy the first factor. Cf. Harrison, 77 Md.
App. at 

(upholding summary judgment when
expert’s only testimony was that "‘[c]ars shouldn’t catch on fire going
down the road’").

   De’Longhi concedes that the Rileys satisfied the second factor —
the occurrence of the accident shortly after the sale. We agree. The
fire occurred only six months after the Rileys bought the heater and
after just three uses. A six-month time span is short enough to satisfy
the second factor. Compare 

, 549 A.2d
                   RILEY v. DE’LONGHI CORPORATION                      7

at 391 (five years after original sale is not a short time), with 

, 484 A.2d at 657 (two or three months is a short
time after sale).

   The third factor is whether the same accident has occurred in simi-
lar products. There is evidence that De’Longhi heaters have been
identified previously as potential fire hazards. In 1991, in cooperation
with the U.S. Consumer Product Safety Commission (CPSC),
De’Longhi offered to provide free replacement control panels on its
oil-filled electric heaters because of the CPSC’s determination that
some of the electrical controls posed a fire hazard. De’Longhi argues
that this evidence is irrelevant because the heaters that were the sub-
ject of the voluntary corrective action were manufactured between
1980 and 1988, and the company concluded that the Rileys’ heater
was manufactured in 1995 or later. The Rileys argue that the date of
manufacture is impossible to determine because the fire damage left
the heater with no identifying marks. All we can say at this point is
that the age of the heater is an open question. Furthermore, Fennell
and Meyer testified that even if the Riley heater was of a newer
design, information about the dangers of older heaters is still relevant
because the layout of the control panel and the type of connectors
used have not changed. We conclude that the CPSC’s determination
that certain De’Longhi heaters made in the 1980s posed a fire hazard
is relevant, at least at this stage. Even if the Rileys’ heater and the
heaters subject to the recall are not exactly the same, there is expert
testimony that the earlier heaters and the one here are sufficiently
similar.

   The fourth factor requires the elimination of other causes of the
accident. Both Fennell and Meyer testified at their depositions that
they did not believe that the fire originated from a source other than
the heater. Fennell basically ruled out an external fire as the cause.
Fennell also disagreed with the conclusion of De’Longhi’s expert,
Wharton, that the extension cord caused the fire. Fennell eliminated
the extension cord because its solid conductors did not show signs of
excessive overheating and there was insulation left on the cord.
Meyer agreed with Fennell, noting that there was no evidence of
extensive burning in the floor area around the extension cord. While
De’Longhi’s expert obviously disagrees with these conclusions, the
Rileys have offered sufficient expert testimony eliminating other
8                  RILEY v. DE’LONGHI CORPORATION

causes of the fire. See 

-89 (children’s tes-
timony that they were not playing with matches and parent’s testi-
mony that the cords were unobstructed was enough at the summary
judgment stage to eliminate other causes of an electric blanket fire).

   Under the fifth factor we consider whether the fire in the Riley
home is the type of accident that does not occur without a defect. The
Rileys argue from common sense: given the circumstances in this
case, the portable heater would not have ignited and caused a fire
unless it was defective. De’Longhi argues that this sort of accident
could have happened in the absence of a defect; specifically,
De’Longhi points to the extension cord as a possible cause of the
accident. It is unclear to us just how the fifth factor should be inter-
preted. On the one hand, we could require the Rileys to show that this
accident would not have happened unless the heater was defective,
even if other causes (such as the extension cord) were not eliminated.
See 

(reading the fifth factor to mean that
"even if other causes are not eliminated, [a plaintiff must show that]
the accident is of a type that does not ordinarily happen unless a
defect exists" (emphasis added)). In that case, the Rileys would not
satisfy this factor because a defect in the heater would not be the only
possible explanation for the fire. On the other hand, if we accepted
Fennell’s and Meyer’s testimony eliminating other causes of the acci-
dent, as we did for factor four, we would inevitably reach the conclu-
sion that the heater could not have caught fire unless it was defective.
See Welge v. Planters Lifesavers Co., 

, 210-11 (7th Cir.
1994) (assuming no consumer misuse, the shattering of a glass jar
does not happen absent a defect); Stackiewicz v. Nissan Motor Corp.,

, 449, 

, 928 (1984) (assuming that other
causes have been eliminated, the malfunction of a product is suffi-
cient evidence of a defect). We do not have to resolve this problem,
however. Even if the Rileys failed to satisfy the fifth factor, their evi-
dence on the first four factors is sufficient to create a circumstantial
case that the heater was defective. See 

(defendants were not entitled to summary judgment when plaintiffs
satisfied three of the Harrison factors).

   In conclusion, the Rileys have satisfied four of the Harrison fac-
tors, and that is adequate in this case to create a genuine issue of
material fact as to whether the heater was defective. In other words,
                   RILEY v. DE’LONGHI CORPORATION                     9

a reasonable jury could draw an inference of a product defect from
the circumstantial evidence that the Rileys have proffered. The district
court therefore erred in granting summary judgment in favor of
De’Longhi. We reverse the judgment and remand the case for further
proceedings.

                                      REVERSED AND REMANDED