Federal Insurance Company ("Federal"), the subrogee of Commonwealth Oil Refining Co., Inc. ("CORCO"), appeals from a judgment of the District Court, for the Southern District of New York (Richard Owen, Judge) awarding Federal $38,610.10, plus prejudgment interest. 587 F.Supp.518 Federal recovered for damage to a portion of a cargo of orthoxylene loaded in June 1977 onto the M.T. Sabine, a vessel belonging to Sabine Towing & Transportation Co., Inc. ("Sabine"). In this maritime action, Federal sought reimbursement for damage to the entire cargo caused by leakage of leaded gasoline from adjacent storage tanks on board the Sabine. The District Court limited damages to the loss of value of the portion of the cargo loaded to the four-foot level, holding that, when loading reached that level, CORCO's loading supervisor had knowledge of facts reasonably warranting cessation of loading to avoid further damage. On appeal, Federal contends that, because CORCO's supervisor was not aware of the lead contamination at the four-foot level, his decision to continue loading was reasonable. Because the evidence permitted the District Court to conclude that CORCO should have ceased loading at the four-foot level, we affirm.

The dispute concerns a cargo of a liquid petroleum product called orthoxylene that was discovered, after loading was completed, to have been contaminated by lead. Orthoxylene is an aromatic hydrocarbon that is used, in its pure form, as a raw material for other chemical products such as insecticides. It is a sophisticated compound, produced according to rigid specifications. CORCO had standard specifications for all customers; among other things, its orthoxylene was required to be a minimum of +25 on the Saybolt color scale,1 to limit non-aromatics to a maximum of .5% of total weight, and to limit the combination of non-aromatics and C9 aromatics to a maximum of 1% of total weight. The cargo's adherence to these specifications was tested before, during, and after loading by analysis of samples. The analyses were performed by E.W. Saybolt and Chas. Martin, both companies licensed to inspect petroleum products.

Departures from the manufacturer's specifications affect the value of orthoxylene to different degrees, depending on the nature of the contaminant and the extent of the deviation from specifications. According to Federal, orthoxylene should contain no lead; any concentration of lead exceeding 50 parts per billion (ppb) would "poison" the substance's catalyst, rendering it unsuitable for its sophisticated uses and thereby greatly diminishing its value. Orthoxylene contaminated with lead is still saleable, though at a greatly reduced price, as a high octane additive for gasoline. Contamination by certain other substances has less dire (and less costly) consequences. A cargo contaminated beyond the specification limit for non-aromatics would be unsuitable for buyers requiring pure orthoxylene, but that cargo could be blended with a sufficient quantity of uncontaminated orthoxylene to bring the mixture within specifications and thus saleable at a premium price. No one contends that blending could be used to bring lead-contaminated orthoxylene back within specifications.

On June 26, 1977, the M.T. Sabine arrived at the CORCO refinery in Guayanilla, Puerto Rico, to receive, among other cargoes, a quantity of orthoxylene. According to Fred Oaks, then superintendent of oil movement and shipping for CORCO and the individual who supervised the loading of the Sabine, the Saybolt inspection company assured him that all shore tanks and feed lines from the tanks to the ship were "on test," i.e., up to specifications. Before loading began, the Sabine's chief officer and a CORCO inspector certified their inspection of the vessel's tanks.

CORCO commenced loading the orthoxylene into the Sabine's No. 2 port, center, and starboard tanks on the evening of June 26. Loading was stopped at the two-foot level so that samples of the cargo could be analyzed to see if the cargo, as it was being loaded, met CORCO's specifications. This analysis, a standard part of loading procedure, revealed that the color readings were low in the port and starboard tanks (+ 21 and + 20) and normal in the center tank (+ 30). Non-aromatics were .39% and C9 aromatics were .49%, both within specifications. No test was made for lead. Oaks, alerted by the low color readings, suspected that some small residue of a contaminant might have been present in the tanks prior to loading. He ordered loading of an additional two feet of cargo to see whether the contaminant that was affecting the color in the port and starboard tanks could be corrected by dilution, a standard practice he had used before.

Loading was halted at the four-foot level so that additional samples could be tested. Both Chas. Martin and E.W. Saybolt did tests at this level. Both analyses showed the color readings unchanged at + 21 port and + 20 starboard. The Martin report noted a non-aromatic weight percent of 1.45 in the port tank and 1.60 in the starboard tank and a C9 weight percent of .92 port and .83 starboard. These figures are roughly three times CORCO's allowance for non-aromatics and nearly 2 1/2 times the specification for the combination of non-aromatics and C9 aromatics. The Saybolt report identified the non-aromatic contaminant as cyclohexane and the C9 aromatic as mercury. Federal does not contend that either of these substances could have caused the continued color variation.

Aware of non-aromatic and C9 aromatic contamination that had not been present at the two-foot level and still unenlightened as to the source of color contamination at both the two- and four-foot levels, Oaks decided to persist with his dilution strategy and resumed loading. A third sampling was taken at the eight-foot level. Saybolt's analysis at this level, a composite of the three storage tanks, revealed a color reading of + 26,2 a non-aromatics percentage of 1.11, and a C9 aromatics percentage of .20. At 1:00 p.m. on June 27, Saybolt notified CORCO that non-aromatic and C9 aromatic levels were high. Loading continued until 3:30 p.m. Only after loading was completed was any test performed to determine whether lead might have contaminated the orthoxylene, presumably by leakage of leaded gasoline from an adjacent storage tank on the vessel. This test showed a level of lead exceeding 400 ppb in all three tanks. At this point, it was decided that the orthoxylene was unfit for its premium use because of lead contamination. The cargo was discharged from the vessel and placed in storage tanks for future sale as a high octane additive for gasoline.

Federal paid CORCO $265,995.65, the difference between the entire cargo's value in its pure form and its reduced value as a gasoline additive. As subrogee of CORCO's claim, Federal then brought suit against the Sabine and her owners for this sum, alleging their responsibility for the contamination of the entire cargo.

The District Court found that Federal was entitled to damages only for the loss caused by contamination of the quantity of orthoxylene loaded up to the four-foot level (1,926 bbls.) and not for the entire cargo (13,468.28 bbls.). In reaching its decision, the Court found that both the discovery of new contaminants and the lack of improvement in color at the four-foot level should have led Oaks to stop loading any more cargo.

Federal's claim is essentially that Sabine, through its negligence, allowed leaded gasoline to leak from tanks adjacent to those receiving the orthoxylene and should therefore be liable for damage to the entire cargo. Sabine responds that, notwithstanding its own initial negligence, Oaks at some point had information that should have caused him to stop loading; any damage to cargo loaded thereafter became CORCO's responsibility rather than Sabine's.3 That point, Sabine contends, was the moment Oaks learned the results of the four-foot analyses.

These conflicting claims are resolved by application of the doctrine of "avoidable consequences." Restatement (Second) of Torts Sec. 918 (1979). Writing for this Court, Judge Friendly authored an oft-cited formulation of the rule:

[A] tort defendant is not liable for consequences preventable by action that reason requires the plaintiff to take.... [T]he community's notions of fair compensation to an injured plaintiff do not include wounds which in a practical sense are self-inflicted.

Ellerman Lines, Ltd. v. The Steamship President Harding, 288 F.2d 288, 290 (2d Cir.1961). Unlike the doctrine of contributory negligence, which traditionally provides a defense to a negligence action, the rule of avoidable consequences is not a defense; it applies "only to the diminution of damages and not to the existence of a cause of action." Restatement (Second) of Torts Sec. 918 comment a. Whereas contributory negligence relates to the use of due care to avoid the event that gives rise to the initial harm, the doctrine of avoidable consequences comes into play at a later stage, denying recovery to the plaintiff for that portion of his loss that could have been avoided by his reasonable efforts. Southport Transit Co. v. Avondale Marine Ways, Inc., 234 F.2d 947, 951-54 (5th Cir.1956).

"[I]n order to prove that a plaintiff's adherence to an initially proper decision how to mitigate damages has become unreasonable, the defendant must show that such adherence had become not merely erroneous but palpably so...." Ellerman Lines, supra, 288 F.2d at 291. The burden of showing that a plaintiff unreasonably failed to minimize damages rests with the wrongdoer. Id. at 290-91; Tennessee Valley Sand & Gravel Co. v. M/V Delta, 598 F.2d 930, 933 (5th Cir.1979). However, if the course of action chosen by the plaintiff was reasonable, the plaintiff can recover despite the existence of another reasonable course of action that would have avoided further damage. Ellerman Lines, supra, 288 F.2d at 290. Therefore, the question in the present case is not whether Oaks chose the path most likely to prevent further damage to the cargo, but whether his decision to continue loading was reasonable under the circumstances.

Upon discovering that the orthoxylene in the port and starboard tanks deviated from the specification as to color, Oaks acted reasonably in deciding to load an additional two feet of cargo in an attempt to dilute the unidentified contaminant. This decision was based on Oaks' past success in using this dilution method to cure initial contaminations and on the fact that it was not unusual for the first portion of a cargo to be somewhat out of specifications because of some foreign substance in the vessel's tanks. Since this initial decision was reasonable, the issue becomes whether Sabine sustained its burden of showing that Oaks' decision to continue loading after receiving the four-foot test results was "not merely erroneous but palpably so." The District Court's decision that Oaks "blundered" by continuing loading beyond the four-foot level was well supported by the evidence.

Oaks' dilution plan could have worked only if the amount of the contaminating substance was fixed. If this were the case, loading more pure orthoxylene would cause a proportionate decrease in the degree of contamination. The results of the four-foot tests, however, revealed that just the opposite was the case. Despite a doubling of the volume of cargo, the color Saybolt readings in the port and starboard tanks were unchanged, indicating that the amount of contaminant was increasing rather than fixed. Moreover, non-aromatic and C9 aromatic contaminants not present at the two-foot level were discovered in the four-foot analyses, again indicating that the contaminants were not fixed but rather were leaking in as the orthoxylene was loaded. The four-foot results demonstrated the futility of any further attempt to bring the cargo within specifications by dilution.

Moreover, the results of the four-foot tests should have convinced Oaks of the need to ascertain the source of contamination before loading more cargo. Federal seems to concede this: "When the cargo was tested, the presence of either a non-aromatic or an aromatic content other than the orthoxylene would alert, warn, and guide those loading ... to a determination of the source of contaminent [sic]" (emphasis added). Appellant's Brief at 4. Although certain contaminants were identified in the four-foot tests, the contaminant causing the out-of-specification color analysis was unidentified and was not being reduced as more pure product was added. There was no claim that cyclohexane or mercury could have caused the low color readings, and the evidence established that excessive amounts of lead could have been the cause. The District Court was entitled to accept the view of Sabine's expert witness and find that the reasonable course of action for Oaks would have been to discontinue loading until the precise cause of the contamination was discovered.

Although Oaks could not reasonably be required to test for all possible contaminants, the record reveals that, had he acted reasonably, he would have tested for lead. Despite the fact that orthoxylene has no standard specification for lead, a sample taken from the dock line prior to loading was tested for lead. And just as Oaks apparently considered the possibility that one contaminant might be cyclohexane because that product was being carried in an adjacent tank (the four-foot tests showed this to be the case), when faced with a contaminant unidentified by either the two- or four-foot tests and not reduced by dilution, he could easily have determined the contents of other adjacent tanks and tested for contamination leaking from them. Oaks clearly should have done so after receiving the test results from the four-foot sample. Such additional testing would have revealed the lead contamination and would have prevented the damage to the rest of the cargo. In fact, this was exactly the reasoning that led Oaks ultimately to test for lead: After loading was concluded, CORCO determined the contents of adjacent tanks and tested the orthoxylene for lead.

Although the contamination of the cargo began because of the ship's negligence in permitting fluid containing lead to leak from an adjacent tank, the ship was not liable for further damage that would have been avoided if the shipper, acting through its agent, had not followed an unreasonable course of conduct. The District Court properly limited recovery to the diminished value of the portion of cargo loaded to the four-foot level, the point at which it became unreasonable for the shipper to load additional cargo without determining whether lead was the source of the contamination.

The judgment of the District Court is affirmed.4