This case causes us to re-examine the definition of seaman as charged in Jones Act cases in light of two recent opinions of the Supreme Court: McDermott Int'l, Inc. v. Wilander, 498 U.S. 337, 111 S. Ct. 807, 112 L. Ed. 2d 866 (1991), and Southwest Marine, Inc. v. Gizoni, --- U.S. ----, 112 S. Ct. 486, 116 L. Ed. 2d 405 (1991) (explicating Wilander ). Specifically, we must determine whether an "employment related connection to a vessel in navigation" must be substantial either in terms of the plaintiff's "permanent assignment to the vessel" or "performance of a substantial part of his work on the vessel." A subsidiary question is whether the district court properly instructed the jury that it could not consider the period of time a vessel was in drydock in determining whether the plaintiff performed "a substantial part of his work on the vessel." We find that, although Wilander overrules the "aid to navigation" element of Salgado v. M.J. Rudolph Corp., 514 F.2d 750, 755 (2d Cir.1975), the "more or less permanent connection" requirement articulated in Salgado remains the law of this circuit. In ignoring this formulation and adopting an alternative formulation, the district court misapplied the law and committed plain error. This error resulted in substantial prejudice to Latsis because the jury was effectively instructed to determine the substantiality of Latsis's connection to the vessel solely on the basis of Latsis's temporal relationship with the vessel. This prejudice was compounded by the district court's further instruction that the time the vessel was in drydock could not be considered--an instruction that improperly served to limit the substantiality of Latsis's connection to the vessel in terms of Latsis's temporal relationship with the vessel. We vacate the judgment of the United States District Court for the Southern District of New York, Loretta Preska, Judge, and remand for a new trial.I. Background

Antonios Latsis brought this action under the Jones Act for compensatory damages for personal injuries sustained by Latsis culminating in the loss of sight to his right eye. These injuries allegedly resulted from the negligence of the ship's doctor who failed to treat promptly Latsis's detached retina. Latsis alleges that he sustained these injuries while employed as a "seaman" aboard a cruise ship then known as the S.S. Galileo, one of a fleet of cruise ships owned and operated by appellees Chandris, Inc., Chandris, S.A. and Trans Oceanic Shipping Co., Ltd. (hereinafter "Chandris").

The injuries are said to have resulted from medical malpractice by the ship's doctor on May 14, 1989, a Sunday on which the Galileo took on passengers at Baltimore, Maryland, for a regularly-scheduled, round trip cruise of five days to Bermuda. Around noon of that Sunday, while the vessel was still in port, Latsis, on board and assigned to the Galileo in his capacity as a supervising engineer for Chandris, began experiencing a central visual disturbance which seemed to originate in the bottom portion of his eye. As the day progressed the obstruction to his vision expanded. Latsis did not appreciate the significance of the problem until the time approached to sail in the late afternoon, when he sought medical guidance from an on board nurse who advised him to see the ship's doctor at once. He did so as the ship pulled from the dock in the port of Baltimore. Once in the ship hospital, he consulted Dr. Gonzalez, the physician employed by Chandris for this vessel. Dr. Gonzalez looked at Latsis's eye through an ophthalmoscope and said it appeared to him to be a detached retina. Gonzalez recommended that Latsis relax until he could see an eye specialist when they arrived two days later in Bermuda. Latsis received no further medical care until after the ship arrived in Bermuda, and no attempt was made to transport him ashore for prompt medical care by means of a pilot vessel, helicopter, or otherwise. During this interval, the impairment of his vision continued to intensify until all central vision in the eye was lost.

The malpractice allegations relate to the fact that Dr. Gonzalez, unlicensed in any state in the United States, neglected to follow the advice of a standard medical textbook which he had on board the vessel which read that "... any patient with a suspected or established retinal detachment should be seen on an emergency basis by an ophthalmologist." When the vessel reached Bermuda, a doctor diagnosed his eye as having a massive detachment of the retina. The doctor recommended immediate hospitalization. After a consultation with Latsis's regular eye doctor in New York and another New York eye specialist, the doctors recommended surgical reattachment of the retina. Although Latsis underwent surgery, the end result was that he lost 75 percent of his central vision and can only see light on the sides of his right eye.

At the time Latsis presented himself for treatment in Bermuda, he was given a Chandris form signed by the ship's captain, chief purser, and Dr. Gonzalez, requesting medical attention be provided to Latsis for his "detected [sic, detached] retina" and describing Latsis as a "crew member" holding the "position" of "supt. engineer."

Evidence in the case indicated that Latsis, a naturalized American citizen of Greek ancestry, has had a lifelong involvement with the sea and shipping beginning with his entry into the Merchant Marine as a cadet in Germany following his graduation from high school, his sailing aboard a merchant ship as a second mate following successful completion of his studies and examinations, and finally with his being appointed captain of a small Mediterranean vessel. Thereafter, he became chief mate on a larger vessel. He came to the United States in 1967 and promptly became a United States citizen. He was employed in San Francisco with a firm of ship surveyors while continuing his education at the University of California at Berkeley. In 1973 he was awarded a Bachelor's Degree of Science and Mechanical Engineering from Berkeley and a year later received a Master's Degree in Naval Architecture, working in the Maritime Industry during this time for the Pacific Marine Association in the ports of San Francisco and Oakland. After graduation he went to New York to work for the American Bureau of Shipping, and later for Southern Star Shipping Company as a superintendent and engineer. He supervised Southern Star's vessels' operations and maintenance for 11 years before becoming employed in 1989 by Chandris out of its Miami, Florida office.

Chandris operated a large fleet of merchant vessels, five of which sailed from Atlantic ports in the United States to the Caribbean, Bermuda, and Puerto Rico. For the first three months, Latsis was employed as a consultant and thereafter became a salaried employee. His duties were not confined to any single vessel but ran to the whole fleet. Latsis's duties included not only overseeing the vessels' engineers aboard ship and, hence, involved a number of voyages, but also planning and directing the maintenance at sea from the shore. While he necessarily familiarized himself with all of the vessels' power plants and engineering functions, he attended several hands-on repairs in a supervising capacity and monitored problems that the ships had or could develop. He was particularly in charge of the electronic equipment, both retrofit and existing equipment.

Latsis's work required him to travel with various ships in the fleet from time to time. The evidence here was conflicting as to how much time he was actually on board ship, Latsis himself testifying that he was on voyages 72 percent of his time and his immediate boss testifying it was more like 10 percent. Latsis maintained his family and residence in Rye, New York, although the company provided him with housing and the use of a rental car in Miami because his office work was regularly there. The vessel on which he was sailing and with which he had more connection than any other vessel in the fleet was the 27-year-old S.S. Galileo, as it was then known, aboard which he had sailed from Miami to Mexico with specialists from Germany to plan for a conversion and upgrading scheduled to take place in Germany in the following year. He helped draw specifications to keep the older ship in repair while sailing on its cruise routes to Mexico and Nassau in the Bahamas. In fact, he had been on board the Galileo at the conclusion of a regular two-day cruise originating from and returning to Baltimore, having joined the ship there on Friday, March 12, two days before the scheduled trip to Bermuda during which his eye injury occurred as above stated. On the Baltimore-Bermuda trip, Latsis was responsible for escorting various shipyard personnel around the ship in connection with the upcoming conversion of the vessel. In fact, he was busy with these people after he first noticed the problem with the vision in his right eye at about noon on Sunday, May 14, up until the time of sailing when he finally went to see the ship's nurse who sent him to Dr. Gonzalez.

According to Demetrius Kaparis, who was a senior vice president of Chandris, Inc. and head of its technical department and technical operations before it was taken over by Celebrity Cruises, Inc., Latsis was hired on a regular basis as a superintendent/engineer with the specific work of supervising the electronic systems of their ships and also upgrading the computerized systems and the computerized communication between the ships and the Chandris Miami office. The company had five superintendent/engineers, two in Miami and three in Greece, and all of the engineering department of the ships and masters of the ships reported to Kaparis for technical matters. The Chandris vessels, depending upon size, carried between 12 and 14 licensed engineers on board, and Kaparis supervised six vessels, all passenger cruising ships customarily calling in at New York, Boston, Philadelphia, Baltimore, Charleston, Wilmington, Miami, and San Juan. Usually, a vessel would be in home port about every seven days. Occasionally, depending on her itinerary, the ship would return to home port after a two or five day cruise. Vessels were usually in home port between 9 and 14 hours. Each ship had a superintendent/engineer, completely responsible, who reported to Kaparis directly. No vessels were assigned to Latsis "because he was taking care of the electronics systems of the ships," in other words, the whole fleet. Latsis reported to Kaparis.

Following Latsis's surgery and recuperation, lasting approximately six weeks, during which he received full pay, he resumed his duties in July, 1989, in Miami, and on September 30 sailed with the Galileo to Bremerhaven, Germany, where the vessel was placed in drydock. The conversion of the vessel was a $65 million job during the course of which the vessel's bottom plates and propeller were removed, the conversion being completed in March, 1990, and the vessel being renamed the Meridian. Latsis, who had been with the vessel the entire time it was in drydock in Bremerhaven, sailed back to the United States aboard the Meridian and continued to work for Chandris until November, 1990, at which time his employment with Chandris terminated. Suit was filed in October, 1991. The district court had subject matter jurisdiction under 28 U.S.C. Sec. 1331 (1988), and 46 U.S.C.App. Sec. 688 (1988).

Trial lasted four days. In the charge to the jury, the judge told the jury that the plaintiff must prove, by a preponderance of the evidence, four elements, namely, that he was a member of the crew of a vessel acting in the course of his employment, that is, that he was a seaman; second, that the defendants were his employer; third, that the defendants or an officer, employee, or agent of one of them was negligent; and, fourth, that such negligence played a part in bringing about the injury to the plaintiff. As to the first element, the jury was told that it could find that the plaintiff was a seaman if the jury found, by a preponderance of the evidence, two things:

First, the plaintiff was either permanently assigned to the vessel or performed a substantial part of his work on the vessel. In determining whether Mr. Latsis performed a substantial part of his work on the vessel, you may not consider the period of time the Galileo was in drydock in Germany, because during that time period she was out of navigation. You may, however, consider the time spent sailing to and from Germany for the conversion. Also, on this first element of being a seaman, seamen do not include land-based workers. (emphasis added).

The second element was stipulated to, and that was that plaintiff's capacity and duties performed "contributed to the function of the vessel or the accomplishment of the vessels' mission or to the operation or maintenance of the vessels during their movements or while at anchor for the vessels' future trips." The court instructed that the second element involved in proving seaman status had been met, and that in considering whether plaintiff has otherwise proved seaman status, "you should know that a vessel in this instance may be a single ship or a number of ships owned and operated by the employer."

Latsis did not object to these instructions in their entirety. Latsis did object, however, to that portion of the charge that "explicitly takes from the jury's consideration the period of time that the Galileo was in drydock in Germany." The jury returned a verdict in favor of Chandris solely on the issue of Latsis's status as a "seaman" pursuant to the Jones Act. On June 17, 1993, the clerk entered a final judgment. On July 15, 1993, Latsis filed a timely notice of appeal. This court has jurisdiction pursuant to 28 U.S.C. Sec. 1291 (1988).

We now hold that the jury instructions were an erroneous application of the law. We therefore remand the case for retrial with jury instructions in accordance with this opinion.

Normally, we will not consider a challenge to a jury charge if a party failed to object at trial. See Fed.R.Civ.P. 51 ("No party may assign as error the giving or the failure to give an instruction unless that party objects thereto before the jury retires to consider its verdict, stating distinctly the matter objected to and the grounds of the objection"). Thus, where "no timely objection was made to the instruction or to the interrogatories in the district court ... we may reverse on this ground only if 'the district court committed plain error in its charge and interrogatories.' " Abou-Khadra v. Mahshie, 4 F.3d 1071, 1078 (2d Cir.1993) (footnote omitted) (quoting Schaafsma v. Morin Vermont Corp., 802 F.2d 629, 636 (2d Cir.1986)). An error is plain "if it results in a miscarriage of justice, or if it is an obvious instance of misapplied law." Abou-Khadra, 4 F.3d at 1078; Air Et Chaleur, S.A. v. Janeway, 757 F.2d 489, 493-94 (2d Cir.1985). Although even a plain error will not warrant reversal if it is harmless, we will reverse where a plainly erroneous instruction misapplies the law as to a core issue in the case resulting in the substantial prejudice of the party challenging the instruction on appeal. Schaafsma, 802 F.2d at 637; see Fed.R.Civ.P. 61. Although Latsis objected to the drydock portion of the instruction, he failed to object to the instruction as a whole. We therefore review Latsis's claim under the plain error standard.

A. The Definition of "Seaman" in the Second Circuit before Wilander

Prior to the Supreme Court's recent opinions in Wilander and Gizoni, this circuit's definition of a "seaman" under the Jones Act was three-pronged:

"the vessel must be in navigation, there must be a more or less permanent connection with the ship, and the worker must be aboard naturally and primarily as an aid to navigation."

Salgado, 514 F.2d at 755 (quoting Klarman v. Santini, 503 F.2d 29, 33 (2d Cir.1974), cert. denied, 419 U.S. 1110, 95 S. Ct. 785, 42 L. Ed. 2d 807 (1975), and Harney v. William M. Moore Bldg. Corp., 359 F.2d 649, 654 (2d Cir.1966)) (emphasis added to distinguish the prongs). For convenience, we denominate these prongs (1) the "vessel in navigation," (2) the "substantial connection," and (3) the "aid to navigation" elements.1 The Second Circuit test very closely followed the test first adopted by the First Circuit in Carumbo v. Cape Cod S.S. Co., 123 F.2d 991, 995 (1st Cir.1941) (citations from sundry circuit courts omitted). Cases in several other circuits also had adopted essentially the same test. See, e.g., McKie v. Diamond Marine Co., 204 F.2d 132, 136 (5th Cir.1953) (defining "member of a crew" in accordance with the three-pronged test); Wilkes v. Mississippi River Sand & Gravel Co., 202 F.2d 383, 388 (6th Cir.) (laborers employed in dredging operations to level off gravel pumped up from river bottom and pumped onto barges held to be seamen under the Jones Act), cert. denied, 346 U.S. 817, 74 S. Ct. 29, 98 L. Ed. 344 (1953); Rackus v. Moore-McCormack Lines, Inc., 85 F. Supp. 185, 187 (E.D.Pa.1949); see also David W. Robertson, A New Approach to Determining Seaman Status, 64 Tex.L.Rev. 79, 94 (1985) (the "Robertson article").

B. The Demise of the "Aid to Navigation" Element--Wilander

Early on, the Fifth Circuit, in Offshore Co. v. Robison, 266 F.2d 769, 779-80 (5th Cir.1959), criticized the "aid to navigation" language of the three-pronged test:

Our review of the cases shows this [aid to navigation] test has been watered down until the words have lost their natural meaning.... With due deference to the Supreme Court, we attach less importance to either of these catchphrases than we do to the cases piled on cases in which recovery is allowed when by no stretch of the imagination can it be said that the claimant had anything to do with navigation and is a member of the ship's company only in the sense that his duties have a connection with the mission or the function of the floatable structure where he was injured.

Id. at 780 (jury question raised whether roughneck as member of drilling crew was a seaman working on main deck of a mobile drilling platform or barge resting on the bottom of the Gulf of Mexico). After review of the Supreme Court and other relevant case law, the Robison court reformulated the criteria for seaman status:

there is an evidentiary basis for a Jones Act case to go to the jury: (1) if there is evidence that the injured workman was assigned permanently to a vessel ... or performed a substantial part of his work on the vessel; and (2) if the capacity in which he was employed or the duties which he performed contributed to the function of the vessel or to the accomplishment of its mission, or to the operation or welfare of the vessel in terms of its maintenance during its movement or during anchorage for its future trips.

Robison, 266 F.2d at 779 (footnote omitted).2 Thus, the Fifth Circuit abandoned the "aid to navigation" element of the common law definition of seaman in favor of language that reflected the tendency of the courts to extend seaman status to employees who do what the Supreme Court would eventually call "the ship's work." See Wilander, 498 U.S. at 355, 111 S.Ct. at 817.

The Robison test differs from the three-pronged test in two other significant respects, however. First, the Robison test dropped the first prong of the test, that is, the "vessel in navigation" element. Second, the Robison test replaced the more or less permanent connection prong of the test with the following requirement:

the injured workman [must have been] assigned permanently to a vessel ... or [must have] performed a substantial part of his work on the vessel.

Robison, 266 F.2d at 779 (footnote omitted).

Eventually, the Robison test became the accepted test of the Fifth Circuit. See, e.g., Balfer v. Mayronne Mud & Chem. Co., 762 F.2d 432, 434 (5th Cir.1985). This court, however, retained the three-pronged test. See, e.g., Salgado, 514 F.2d at 755; Buccellato v. City of New York, 808 F. Supp. 967, 971 (E.D.N.Y.1992). In Johnson v. John F. Beasley Constr. Co., 742 F.2d 1054, 1062-63 (7th Cir.1984), cert. denied, 469 U.S. 1211, 105 S. Ct. 1180, 84 L. Ed. 2d 328 (1985), the Seventh Circuit attempted to highlight the difference between the two tests, focussing on the "aid to navigation" prong of the three-pronged test and the significant contribution aspect of the Robison test. Reasoning that the significant contribution aspect of the Robison test was "too broad," the Seventh Circuit adopted yet another test for seaman status--one combining the "more or less permanent connection" and the "vessel in navigation" elements of the three-pronged test and a requirement that "the person injured made a significant contribution to the maintenance, operation, or welfare of the transportation function of the vessel." Id. at 1063.

The Supreme Court in Wilander addressed the conflict between the Fifth Circuit's Robison test and the Seventh Circuit's Johnson test, Wilander, 498 U.S. at 340, 111 S.Ct. at 809-10, but only as to one difference between the two tests--the aid to navigation requirement. The Supreme Court jettisoned the "aid to navigation" element, and announced that "the better rule is to define 'master or member of a crew' under the [Longshore and Harbor Workers Compensation Act], and therefore 'seaman' under the Jones Act, solely in terms of the employee's connection to a vessel in navigation." Wilander, 498 U.S. at 353, 354, 111 S. Ct. at 816-17. The Court observed, that "[t]his rule best explains our case law and is consistent with the pre-Jones Act interpretation of 'seaman' and Congress' land-based/sea-based distinction." Id. at 354, 111 S. Ct. at 816-17. The Court continued to write, "It is not the employee's particular job that is determinative, but the employee's connection to a vessel," id., and concluded that "[t]he key to seaman status is employment-related connection to a vessel in navigation." Id. at 355, 111 S. Ct. at 817.

In so doing, the Court adopted only so much of the Robison test as required that a seaman's duties " 'contribut[e] to the function of the vessel or to the accomplishment of its mission.' " Wilander, 498 U.S. at 355, 111 S.Ct. at 817 (quoting Robison, 266 F.2d at 779). The Court did not adopt the Robison test in its entirety. Indeed, the Court cited, with apparent approval, the Robertson article which called for revision of the Robison test, stating that "[a]ll who work at sea in the service of a ship face those particular perils to which the protection of maritime law, statutory as well as decisional, is directed." Id. As we discuss more fully below, the Court did not follow Robison's departure from the "vessel in navigation" element of Salgado. On the contrary, the Court's language, "[t]he key to seaman status is employment-related connection to a vessel in navigation," id. (emphasis added), confirms that this element remains an important part of the definition of a seaman. Also, the Court expressly noted that, "[w]e are not called upon here to define this connection in all details." Id. Further, the Court in Wilander did not state that an employment-related connection to a vessel was a complete definition of seaman. See id. 498 U.S. at 355, 111 S. Ct. at 817-18 (stating that "we hold that a necessary element of the connection is that a seaman perform the work of a vessel") (emphasis added). The Court offered no indication as to whether the substantiality of the connection must still be established or, if so, whether it would be determined pursuant to the "more or less permanent connection" element of the "Johnson test" (as well as the three-pronged Salgado test) or pursuant to the language of the two-part Robison test.

C. The "Substantial Connection" Element after Wilander

As discussed above, the Supreme Court in Wilander did not determine under what circumstances an employment related connection with a vessel in navigation would be sufficiently substantial to qualify a plaintiff for seaman status. It is clear, however, that the Court did not express a preference for the criteria contained in the Robison test. Left unclear, therefore, is just what sort of connection will amount to a substantial connection. Indeed, the Supreme Court did not even recognize, much less resolve the conflict between the criteria used to determine substantial connection under the Robison test and those used under the Johnson and Salgado tests. Again as noted above, the Court limited its holding to the "aid to navigation" element of the three-prong test, concluding that "[t]he key to seaman status is employment-related connection to a vessel in navigation," but declined "to define this connection in all details." 498 U.S. at 355, 111 S. Ct. at 817.

Nor does the Supreme Court's opinion in Gizoni offer much guidance in the way of what constitutes a substantial connection. In Gizoni, decided after Wilander, the Court held that a maritime worker whose occupation is one of those enumerated in the LHWCA may be nevertheless a seaman within the meaning of the Jones Act, at least where the harbor worker is also a "member of a crew of any vessel," as explained in Wilander. Gizoni, --- U.S. at ----, 112 S. Ct. at 491. The Gizoni Court quoted Wilander in saying that " '[t]he key to seaman status is employment-related connection to a vessel in navigation.... It is not necessary that a seaman aid in navigation or contribute to the transportation of the vessel, but a seaman must be doing the ship's work.' " Id. --- U.S. ----, at 492 (quoting Wilander, 498 U.S. at 355, 111 S.Ct. at 817). No mention is made in either Wilander or Gizoni of the "permanent connection" language in the Robison case.

In this case, there is no question that Latsis at the time of his injury, was connected to a vessel in navigation, that his connection was employment-related, and that at least at the time of his injury he was "doing the ship's work," namely, acting as a supervising engineer on board the vessel. In fact, it is precisely this part of the jury charge to which Chandris stipulated. Whether this connection was substantial, however, was in dispute. In charging the jury on this question, the district court appears to have adopted language from the two-part Robison test, instructing the jury that to find Latsis a seaman they must find that:

First, the plaintiff was either permanently assigned to a vessel or performed a substantial part of his work on the vessel.

See Robison, 266 F.2d at 779 ("the injured workman [must have been] assigned permanently to a vessel ... or [must have] performed a substantial part of his work on the vessel.") (footnote omitted). Absent disapproval by the Supreme Court, the "more or less permanent connection" formulation of the substantial connection requirement remains the law of this circuit. Therefore, it was plain error for the district court to ignore this formulation and adopt the formulation of the Fifth Circuit in Robison.

The formulation of substantial connection followed by this circuit in Salgado is quite different from that followed by the Fifth Circuit in Robison. The more or less permanent connection element of the Salgado test is not merely a temporal concept. Under Salgado, the substantial connection element is satisfied where there is a "more or less permanent connection with the ship." See, e.g., Salgado, 514 F.2d at 755 (internal quotation marks omitted). Later cases expanded the "substantial connection" element to include connections not just to one particular ship, but to an identifiable group of vessels.3 See Magnolia Towing Co. v. Pace, 378 F.2d 12, 13 (5th Cir.1967) (per curiam) (pilot connected with several of defendant's tugboats held a seaman); Braniff v. Jackson Ave.-Gretna Ferry, Inc., 280 F.2d 523, 528 (5th Cir.1960) (holding that a master mechanic and an assistant who performed all maintenance, repair and overhaul work on several of defendant's ferries, often while ferries were operating, could be seamen); see also Guidry v. Continental Oil Co., 640 F.2d 523, 528 n. 14 (5th Cir.) ("relationship between the individual and an identifiable vessel or group of vessels must be substantial in point and time, not spasmodic") (emphasis added), cert. denied, 454 U.S. 818, 102 S. Ct. 96, 70 L. Ed. 2d 87 (1981); Rotolo v. Halliburton Co., 317 F.2d 9, 12-13 (5th Cir.) (injured person must have been assigned to or performed a substantial part of his work on "a particular vessel or several specified vessels.") (citing Braniff ), cert. denied, 375 U.S. 852, 84 S. Ct. 111, 11 L. Ed. 2d 79 (1963).4

As pointed out by Judge Glasser in Buccellato, the substantial connection requirement has not been interpreted by the courts of this circuit in a narrow temporal sense. Buccellato, 808 F. Supp. at 972. As Judge Glasser observed, Salgado, quoting two earlier decisions, Klarman and Harney, only required that plaintiff have a "more or less permanent connection with the ship." Salgado, 514 F.2d at 755 (emphasis added). The court in Klarman had noted that the permanent connection requirement has not been read strictly but meant that the connection must be more than "temporary or transitory." Klarman, 503 F.2d at 33 (quoting Bullis v. Twentieth Century-Fox Film Corp., 474 F.2d 392, 394 (9th Cir.1973) ("Significantly, the connection may be less than permanent, but it cannot be temporary or transitory")). In Harney, the court noted, but did not decide, that the permanent connection requirement may include "shore-based employees who, though regularly coming aboard, spend less than a full shift there." Harney, 359 F.2d at 654.

An earlier Second Circuit case, not mentioned by Judge Glasser, is even more supportive of his conclusions in Buccellato that "the case law in this Circuit does not unequivocally require a Jones Act seaman to be substantially connected to a vessel or group of vessels as opposed to being connected to a vessel or group of vessels on a steady basis." Buccellato, 808 F. Supp. at 972. That is the case of Weiss v. Cent. R.R. Co. of New Jersey, 235 F.2d 309 (2d Cir.1956). There, a plaintiff who in the course of his 19-day employment worked seven days on ferry boats as a deckhand or wheelsman and on six other days was employed on shore as a doorman or bridgeman, jobs connected with the docking and departure of ferries, was held a Jones Act seaman. The court noted that weighing against his claim were the factors that he was employed aboard ship for only a short period of time, that during that period he slept and ate most of his meals ashore, that he was paid an hourly wage and worked for the most part an eight-hour day, that he was employed for part of the time on jobs ashore, that he was working on an extra-man rather than on a regular basis, and that he had no seaman's papers. 235 F.2d at 311. Further, the Weiss Court cited Bailey v. City of New York, 55 F. Supp. 699 (S.D.N.Y.1944), affirmed, 153 F.2d 427 (2d Cir.1946), which held that a ferry boat engineer who worked a regular eight-hour day and slept at home was a seaman, as was a similar land-based dredge employee in Pariser v. City of New York, 146 F.2d 431 (2d Cir.1945). See Weiss, 235 F.2d at 3132.5

Unlike the Salgado formulation of the substantial connection requirement, the language of the Robison test frames the substantial connection issue primarily, if not solely, in temporal terms. Specifically, the Robison test provides that:

there is an evidentiary basis for a Jones Act case to go to the jury ... if there is evidence that the injured workman was assigned permanently to a vessel ... or performed a substantial part of his work on the vessel.

Robison, 266 F.2d at 779 (footnote omitted). Early cases applying this test recognized that "the word 'permanent' has never been assigned a literal interpretation under the Jones Act and should not be given a 'wooden application,' but rather, is to be used as an analytical starting point instead of a self-executing formula." Guidry, 640 F.2d at 529 (citations omitted). Through subsequent applications of this test, however, the Fifth Circuit clearly expressed its view that substantial connection is a temporal concept to be determined by the duration of the putative seaman's connection to the vessel. Barrett v. Chevron, U.S.A., Inc., 781 F.2d 1067, 1074-75 (5th Cir.1986) (en banc); Longmire v. Sea Drilling Corp., 610 F.2d 1342, 1347 & n. 6 (5th Cir.1980).6

By adopting the Robison test for a substantial connection, the district court limited the jury's consideration to Latsis's temporal relation to the vessel. This substantially prejudiced Latsis. Had the jury not been instructed to determine whether Latsis's contribution to the vessel was substantial only in temporal terms of a "permanent assignment" or having performed a substantial part of his work aboard the vessel, the jury could well have found Latsis to be a seaman. A jury could find that he did have a substantial connection with the vessels though he may not have been on board ship a great percentage of his time because he did the specific work of supervising the electronics systems of the ships, upgrading their computerized systems, and dealing with the computerized communication between the ships and office. Thus, a jury could have found a substantial connection even though Latsis's work did not involve "permanent assignment" to a vessel or to a fleet of vessels and even if we take, as the jury might have taken, Kaparis's testimony that Latsis was on board ships only 10 percent of the time at face value. We therefore hold that the district court's instruction erroneously applied the law of this circuit regarding a core element of Latsis's claim--to the substantial prejudice of Latsis.

D. The "Drydock Instruction"

We turn now to the district court's "drydock instruction." As noted above, the district court instructed the jury:

In determining whether Mr. Latsis performed a substantial part of his work on the vessel, you may not consider the period of time the Galileo was in drydock in Germany, because during that time period she was out of navigation.

In so instructing, the district court seems to have injected the third prong of the old Salgado test--"vessel in navigation," into the substantial connection requirement. We find that this instruction further emphasized the putative seaman's temporal connection to the vessel at the risk of confusing the jury and frustrating the underlying policy of the Jones Act.

Even before the enactment of the Jones Act in 1920, Learned Hand, in extending seaman status to a bartender on board a vessel, remarked that there is "in principle no reason why there should be an artificial limitation of rights to those engaged in the navigation of the ship, to the exclusion of others who equally further the purposes of her voyage." The J.S. Warden, 175 F. 314, 315 (S.D.N.Y.1910). Judge Hand's words capture the spirit of the Jones Act, to compensate those injured in the course of work exposing them to the perils of the sea. This spirit is captured in the "vessel in navigation" element. Moreover, this element of the three-pronged test for seaman has survived Wilander. See Wilander, 498 U.S. at 355, 111 S.Ct. at 817-18 ("the key to seaman status is employment-related connection to a vessel in navigation") (emphasis added).

Courts have interpreted "perils of the sea" broadly. Not only may seamen recover under the Jones Act for injuries traditionally associated with the sea--a swinging boom, a sinking ship--but seamen may recover for injuries more commonly experienced by land-based employees. See, e.g., Pace, 378 F.2d at 12 (recovery allowed for seaman injured in automobile accident while en route to vessel). In this case, Latsis's employment did expose him to the perils of the sea--in fact, his injury was the result of such a peril in the sense that while on board a seaman is very much reliant upon and in the care of the ship's physician. If that physician is unqualified or engages in medical malpractice, it is just as much a peril to the mariner on board as the killer wave, the gale or hurricane, or other dangers of the calling. Not every worker connected with a vessel is exposed to the perils of the sea, however. On the contrary, only workers connected to a "vessel in navigation " are so exposed--even if the exposure results in nothing more than an injury sustained while driving to the vessel. See id. The "vessel in navigation" element, therefore, does not limit the substantiality of an employee's connection with a vessel. Rather, the "vessel in navigation" element was designed to limit seaman status to those whose employment regularly exposes them to the hazards of the sea.

We note that the case upon which the district court evidently relied, in instructing the jury that in considering the substantial connection element it should discount the time during which the Meridian, formerly the S.S. Galileo, was in drydock in Germany, involved an accident while the vessel was in drydock and, hence, not in navigation. See McKinley v. All Alaskan Seafoods, Inc., 980 F.2d 567, 569-70 (9th Cir.1992). McKinley may properly be read as holding that a ship undergoing complete conversion, as the vessel was in that case, takes a vessel out of navigation. See West v. United States, 361 U.S. 118, 121, 80 S. Ct. 189, 192, 4 L. Ed. 2d 161 (1959) (per curiam) (non-Jones Act case finding no warranty of seaworthiness while vessel was undergoing major structural repairs). But this does not necessarily mean that the time during which a vessel is undergoing conversion cannot count towards the substantiality of an employment-related connection with the vessel. During the course of Latsis's employment with Chandris--omitting the consulting time--from approximately March of 1989 until his termination in November, 1990, he spent some six months working seven days a week on board the vessel, supervising the project as the only naval architect, including the drydocking, bottom removal, cleaning and removal of the propellers, overhauling the main engines, rebuilding the boilers, and reconstructing the accommodation spaces and public rooms. Why this time should not count in terms of determining the substantiality of his connection to the Chandris vessels, we do not know. While we express no opinion as to whether Latsis would have been covered had he been injured while the vessel was in drydock in Bremerhaven, we do suggest that in terms of meeting any requirement of connection to the vessel or vessels, this time can count.

Although Salgado's formulation of the substantial connection element remains the law of this circuit after Wilander and Gizoni, we think it useful to clarify this formulation in light of Wilander and Gizoni. Such clarification is made all the more pressing by the fact that although Wilander and Gizoni left Salgado's formulation of the substantial connection requirement intact, the Supreme Court in those cases did jettison another aspect--the aid to navigation element--of that test.

Clarification is also important considering that the Fifth Circuit has continued to adhere to its narrow interpretation of the substantial connection requirement. In Bach v. Trident Steamship Company, Inc., 920 F.2d 322 (5th Cir.), vacated, 500 U.S. 949, 111 S. Ct. 2253, 114 L. Ed. 2d 706 (1991), two members of the panel held that a compulsory pilot was not a seaman because he was not connected to a vessel or an identifiable fleet of vessels, and refused to find an exception in favor of a maritime worker who spends virtually all of his time performing traditional seaman's duties--work closely related to the movement of vessels--but does his work on short voyages aboard a large number of vessels. Judge Brown dissented, stating that Congress did not mean to exclude from the protection of the Jones Act compulsory pilots while aboard in directing the navigation of a vessel. See Bach, 920 F.2d at 327-33 (Brown, J., dissenting). The Supreme Court vacated the opinion and remanded for reconsideration in light of Wilander. 500 U.S. at 949, 111 S.Ct. at 2253. The panel majority reinstated its decision, holding that it had previously "concluded that [the pilot] was not a seaman because he was not permanently assigned to any particular vessel or fleet of vessels." Bach v. Trident S.S. Co., Inc., 947 F.2d 1290, 1291 (5th Cir.1991), cert. denied, --- U.S. ----, 112 S. Ct. 1996, 118 L. Ed. 2d 592 (1992). Judge Brown again dissented, stating that "although Wilander turns out not to be the decisive guide," what it does say affords "undeniable support" to the views he had set forth in his initial dissent. Bach, 947 F.2d at 1291 (Brown, J., dissenting). In addition, he quoted the Wilander case extensively and we may read him as saying that at least in the case of a pilot and his unique status as a seagoing functionary, the requirement of "permanency" set forth in the various cases as a part of the test is simply no longer of force and effect, at least in the temporal sense.

The substantial connection requirement has two purposes. First, by requiring a "more or less permanent connection with a vessel " or with a fleet of vessels, it excludes from Jones Act seaman status workers whose work may bring them into contact with many vessels but who are not connected with any particular vessel or identifiable group of vessels. Through this exclusion, the substantial connection element leaves workers with no connection to a particular vessel or group of vessels to coverage under the Longshore and Harbor Workers' Compensation Act (LHWCA)--from which "seaman" or, more precisely, "a master or member of a crew of any vessel" are themselves excluded. 33 U.S.C. Sec. 901 et seq. (1988). Second, by requiring a "more or less permanent connection with a vessel" or fleet of vessels, it excludes from coverage those whose contribution, although limited to a single vessel or fleet of vessels, was nevertheless only a temporary or transitory connection with the vessel. See Klarman, 503 F.2d at 33 (putative seaman's connection must be more than "temporary or transitory").

It seems that much of the confusion surrounding "more or less permanent connection with a vessel" language is attributable to the fact that the language is designed to perform two functions--i.e., exclude two different types of workers from seaman status. See Buccellato, 808 F. Supp. at 972. Indeed, the Fifth Circuit's adherence to a narrowly interpreted permanent assignment requirement has been criticized by the Supreme Court of Louisiana on essentially this ground. Folse v. W. Atlas Int'l, Inc./Downhole Seismic Servs., 593 So. 2d 341 (La.1992). In Folse, the plaintiff had been held by the Louisiana Court of Appeal not to be a seaman because he was not permanently attached to a specific vessel or an identifiable fleet of vessels. See Folse v. W. Atlas Int'l, Inc., 580 So. 2d 482 (La.App. 4th Cir.1991), rev'd, 593 So. 2d 341 (La.1992). In reversing, the Supreme Court of Louisiana discussed the permanent attachment language of Robison and Bach, saying that "[t]his requirement was designed to distinguish between transitory, land-based employees and those working aboard a vessel for a voyage." 593 So. 2d at 343 (citing the Robertson article). The majority pointed out that Wilander did not mention more or less permanent attachment to a vessel or fleet of vessels,7 and that Gizoni cast further doubt on this requirement because Gizoni "only 'rode' the platforms during working hours." Id. The Louisiana court went on to say that Folse was not a land-based worker providing shore services to docked vessels but rather a seaman who went to sea on specific short voyages and who was not merely a transitory employee. The court went on to say that although Folse's assignment was less transient than Bach's, it was not permanent. Rather, "[i]t involved substantial work on given missions." Id.8

We find that much confusion could be alleviated by breaking down the "more or less permanent connection with a vessel" language into two separate elements. Thus, the substantial connection requirement is met where the putative seaman establishes an employment-related contribution that is (1) limited to a single vessel or group of vessels and (2) substantial in terms of its (a) duration and (b) nature. In considering whether the putative seaman's contribution is substantial in terms of its duration, the jury may consider the number and frequency of similar contributions to that vessel or to an identifiable fleet of vessels. Further, in considering whether the contribution is substantial in terms of its nature, the jury may consider the amount of time and effort spent ashore preparing for the contribution, as well as the amount and time and effort spent in drydock consummating the contribution.

We reiterate that after Wilander and Gizoni the test of seaman status under the Jones Act is an employment-related connection to a vessel in navigation. The test will be met where a jury finds that (1) the plaintiff contributed to the function of, or helped accomplish the mission of, a vessel; (2) the plaintiff's contribution was limited to a particular vessel or identifiable group of vessels; (3) the plaintiff's contribution was substantial in terms of its (a) duration or (b) nature; and (4) the course of the plaintiff's employment regularly exposed the plaintiff to the hazards of the sea.

In this case, the first of the foregoing criteria was stipulated to. As to the second, there is no dispute in the evidence and the verdict should accordingly be directed. The third--the substantiality element--is not only in dispute, but the evidence is not clear as to what extent Latsis's land-based and drydock work bore on (3)(b). Therefore, we instruct the trial court to frame its instructions to the jury in light of Wilander and Gizoni and other cases and opinions above referred to, not in terms of assignment to a vessel or as requiring permanency but, rather, in terms of the employment-related connection to the Chandris fleet. The jury also should be instructed that in determining what is "substantial" in terms of the duration of Latsis's contribution to the vessel or fleet of vessels, the number and frequency of similar contributions may be taken into account. Further, the jury should also be instructed that in determining what is "substantial" in terms of the nature of Latsis's contribution, Latsis's on-shore preparation should be taken into account, as well as the time Latsis spent in drydock in Bremerhaven, Germany consummating his contribution. Finally, the jury should also be instructed to determine whether the course of Latsis's employment regularly exposed him to the hazards of the sea. As to the proof of these matters, the putative seaman of course has the burden.

Judgment vacated, case remanded for retrial with instructions to the jury in accordance with this opinion.

I respectfully dissent.

Though I agree that the four criteria articulated by the majority ante at 57 should govern the determination of whether an employee is a "seaman" within the meaning of the Jones Act, even under that formulation I am unpersuaded that the non-dry-dock part of the district court's instructions, to which Latsis made no objection, constituted "plain" error. Further, I am unpersuaded that the district judge erred in instructing the jury that, in determining whether plaintiff Latsis was a seaman, it should disregard the period during which the Galileo was in dry dock. The Jones Act concept of " 'seaman' " contains a "land-based/sea-based distinction." McDermott International, Inc. v. Wilander, 498 U.S. 337, 354, 111 S. Ct. 807, 817, 112 L. Ed. 2d 866 (1991). The 65 million-dollar dry-dock conversion and reconstruction of the Galileo was unquestionably a land-based activity. In my view, this part of the instructions--the only part to which Latsis objected--was correct.

Accordingly, since in my view the only portion of the instructions to which Latsis objected was not erroneous, and since the remainder of the instructions did not constitute plain error, I would affirm.