Reproduced with permission of 27 American Journal of Comparative Law (1979) 291-310
[This is a commentary on Passing of Risk under the 1978 Draft Convention which, with modifications at the 1980 Vienna Diplomatic Conference, became the CISG. The 1978 Draft provisions discussed are: Article 78 (became CISG Article 66), Article 79 (became CISG Article 67), Article 80 (became CISG Article 68), Article 81 (became CISG Article 69), and Article 82 (became CISG Article 70). For comparison of each of these provisions of the 1978 Draft with the CISG language that emerged, click on the links identified above.]
The Special Commission appointed by the Hague Conference in 1951 to propose a revised draft of the Uniform Law on the International Sale of Goods reported that the subject of risk was "one of the most serious problems which were presented to the draftsmen."  The difficulty of preparing uniform rules on risk lies in the inherent nature of the concept of risk itself, and the diversity of its treatment under municipal laws. The provisions governing risk in the law of sale are of considerable practical importance. While the goods in a modern sales transaction will often be insured, the rules of risk determine the practical obligations to take out insurance, to press a claim against the insurer or bear the burden of inadequate cover, and which party is concerned to salvage damaged goods. "Risk" as a legal concept refers to accidental injury to the goods. It therefore covers theft, seizure, destruction, damage and deterioration. Which kind of injury is concerned can be significant. In the first three cases, the goods are totally lost. If they are damaged, there may be a possibility of repair. And deterioration when goods are perishable may well have resulted from delay in performance of the contract, so the law might choose to distinguish this particular risk as resting on the party in default.
The central question in the consideration of risk is one of time: when does the risk pass from the seller to the buyer? The answer will affect the obligation to pay the price. Despite loss or damage, the buyer must pay if the risk at the time of injury was on him; not so if it was still on the seller. However, risk concerns not merely price risk, but can have a wider effect. If the goods perish while at the seller's risk, not only is the buyer not liable for the price, but the seller may be liable for damages for non-delivery (e.g., replacement at a higher price). Equally, if the risk has passed to the buyer, not only is he liable for the price, but the seller may be able to recover damages for non-acceptance (e.g., storage charges). In some legal systems, only frustration of the contract would preclude such liability in damages, and this is an aspect of the distinction between risk and frustration. Moreover, the transfer of risk will be affected by a material breach of contract. Non-conformity of the goods justifying their rejection, or delay by the buyer preventing their delivery, may well alter the allocation of risk. Finally, where the transaction is a documentary sale -- a frequent occurence in international commerce -- a tender of improper documents can provoke further intricacies. Thus risk is an elusive concept, passing in the abstract from seller to buyer and sometimes back again, intimately connected with the parties' obligations, and embracing several notions which must occasionally, but cannot always, be separated.
Problems of accidental loss and of insurance cover are of course found throughout the world. While the various municipal laws display broad similarities in the practical effect of their rules on risk, there are many significant divergences of detail. Moreover, even where the substantive result is the same, the theoretical rules which produce it are frequently discrepant. Whether transfer of risk is linked to delivery of the goods is one basic example. The time of delivery determines the passing of risk in the Austrian, German and Scandinavian laws of sale, and in the UCC, but it has no such critical significance in French law or in the English Sale of Goods Act (which has been followed in much of the Common law world), where risk is tied to the passing of property. And in Spanish law -- and thus the laws of some South American countries  -- risk passes on delivery in a mercantile sale, but generally as early as the formation of the contract in a private transaction. This distinction between commercial and non-commercial sales is found in a number of systems. Another dichotomy is between specific and unascertained goods; there are special provisions dealing with these in many laws of sale, but in Japanese law the specification of the goods is a notion which provides the very foundation of the rules on risk 
Insofar as the drafting of a uniform law involves an effort at compromise, this variety among national laws makes the task daunting. However, unification is not simply a synthetic exercise, drawing on the different legal systems to develop concepts acceptable to all, but involves innovation as well, seeking rules which will most effectively facilitate modern international trade. There are thus two basic issues which confront the draftsmen of uniform rules on the passing of risk. The first is one of substance: What practical policies should the provisions concerning risk seek to achieve; which party, in a given case, should bear the risk? Secondly, there is a question of form. How can the articles best be drafted in order to ensure that these policies are realized effectively, coping with the immense range of factual situations which may arise, without resorting to undue technicality?
Commercial men seek in the law a body of rules which is both reasonable and yields a result ascertainable with certainty. This is especially true of the passage of risk. Even if the role of the law of sale in general is to give effect to the parties' intention on contracting, risk is not a matter they are likely to have considered if there is no express term to this effect, for casualty is rare and goods are usually insured. There is therefore a clear need for a practical orientation in the legal provisions concerning risk. One criticism of ULIS was that in resolving the differences between national laws, too little attention was paid to commercial practices. The rules on risk in municipal legislation may be seriously outdated. Under English law, for example, in order to achieve a satisfactory result, the courts frequently have to displace the statutory provision linking the passing of risk to property by inferring a contrary intention of the parties. When the law must regularly resort to judge-made rules to determine speculative, and possibly non-existent states of mind, the divergence between code or statute and commercial reality has become too great. On the question of risk, as much as in any other aspect of sales, the uniform law should abandon what Rabel once called the "awesome relics from the dead past."  It is not surprising that in the long process of drafting ULIS it was early on accepted that risk should be separated from property. Indeed, the 1978 UNCITRAL Draft Convention has maintained the policy of ULIS in rejecting the property concept as a tool for deciding the various specific issues which arise in sales transactions.
Turning to the issue of form, the degree of detail the law should display points to the basic contrast in style between Common and Civil law legislation. In its provisions on risk, the UCC in particular is notable for its elaborate and specific treatment of the various issues, whereas the Civil codes employ much more general expressions. The draftsmen of a uniform law must consider, for example, whether it is helpful to deal expressly with the effect of the seller's breach of contract upon the passage of risk, or sufficient to leave this to be inferred from the buyer's remedies. ULIS is relatively long, for which it has been criticized. In drafting the UNCITRAL Convention, a deliberate effort was made to reduce the number of articles. And if brevity in a uniform law is of value, clarity is at a premium. The law has to be translated into different languages and be applied in the context of widely varying legal systems. Thus complicating distinctions, such as separate rules for merchants and non-merchants or for commercial and non-commercial sales, have no place in the absence of some compelling reason. Ambiguous terms should not be employed and, above all, the law should be founded on concrete steps in the performance of the contract by one party or the other. Lack of clarity is one of the major criticisms directed at ULIS, and is thus one of the important criteria by which its proposed successor is to be judged.
ULIS and the Draft Convention
The section on risk comprises the final chapter of ULIS, and it has retained this position in the 1978 UNCITRAL Draft Convention. The opening article of this chapter in the Draft Convention follows ULIS also in expressing the essential nature of risk as an exception to the synallagma, the mutually interdependent obligations of the seller to make delivery and the buyer to pay the price. Art. 78 declares that when risk has passed to the buyer, he shall pay the price despite loss or damage to the goods, unless this was due to an act or omission of the seller.
Under ULIS, the basic rule on risk is stated in art. 97(1): risk passes on "delivery . . . in accordance with the contract and the present law." The concept of "delivery" (délivrance) thus plays a pivotal role in the passing of risk. It is defined in art. 19, and means the handing over of conforming goods. The unwieldy nature of this notion rapidly becomes apparent. It is not delivery in the English sense, for if the goods are defective there will be no délivrance. Nor does it amount to performance by the seller of his part of the contract, for if the buyer refuses to accept conforming goods there will be no handing over and so again no délivrance. The role of délivrance in ULIS has been appropriately called fundamental, for it governs not only the question of risk, but determines also the buyer's obligation to pay the price. Yet there is no English word which adequately expresses the meaning of the French term, and it has proved difficult to translate into other languages as well. Moreover, the employment of a single legal concept to perform two different substantive tasks leads to exceptions being required in the treatment of both. Regarding risk, délivrance proves an inadequate device on the handing over of defective goods, as the seller would retain the risk indefinitely even if the buyer kept the goods, so a special provision is required to provide a different principle in this case. As délivrance alone would not produce a satisfactory solution upon an improper shipment (e.g., where the seller does not conclude a proper contract of carriage), the additional phrase, "in accordance with the provisions of the contract," is required in art. 97(1). Having decided to rest the transfer of risk on deliverance, the law then had to make exceptions and additions to avoid many of the consequences.
The comments submitted to UNCITRAL by the Governments and their representatives echoed the earlier criticism, and the Working Group redrafting the law, after considerable discussion, decided to separate risk from "delivery," and to develop rules for the former based on overt commercial events. This left the concept of delivery free to cope with other tasks, while the rules on risk would no longer depend upon the interrelation of widely dispersed articles. The substantive provisions on risk are accordingly brought together in arts. 79-82. They can most conveniently be approached by considering separately three categories: (1) where there is no breach of contract; (2) where there is a breach by the seller; (3) where there is a breach by the buyer.
An international sale frequently involves intermediate carriage by a third party. If loss or damage occurs during transit, it will be the buyer who discovers this when the goods reach their destination (unless the seller himself has an obligation to deliver there). The buyer -- at least in a commercial transaction -- will accordingly be in a better position to press a claim against the carrier or insurer and to salvage the goods by repair or sale. As a matter of policy therefore, the transit risk should ordinarily be on him. This will, in addition, prevent the buyer using minor damage as a pretext for rejecting the goods when the market has declined and he is anxious to escape from a bad bargain.
The UNCITRAL Draft Convention achieves this result, with an evident gain in clarity over ULIS where reference has to be made to a sub-rule of délivrance. Under the Draft Convention, a shipment contract (i.e., where the seller does not have to hand over the goods to the buyer at destination) of specific goods is dealt with in a separate provision: art. 79(1). Risk passes "when the goods are handed over to the first carrier for transmission to the buyer." And the article makes clear that "handing over" refers to physical possession of the goods, not legal control over them, stating that the seller's retention of documents controlling the goods is immaterial. This is of obvious importance in a documentary sale. The Draft Convention also introduces a qualification that if the seller is required to hand over the goods to a carrier "at a particular place other than the destination," risk will only pass then. The purpose of this is said to be to ensure that risk will only pass at the first seaport, not on handing over to an inland carrier. Thus if a seller in Detroit makes a contract CIF Antwerp, shipment from New York, risk would not pass when the goods are handed over to the railroad in Detroit. While this provision makes clear that a seller remains responsible for the goods if he uses his own trucks for inland carriage, with containerized transport risk should probably pass at the start of inland carriage by an independent carrier at least, for it might be impossible to ascertain the moment of damage after the goods are packed. Perhaps in such a case the parties would specify this in their contract; one must hope so, for the 1978 Draft Convention has not grappled with the legal problems caused by the growing use of containers.
In the normal case of a shipment of unascertained goods, ULIS creates particular confusion. As well as handing over the goods to the carrier, the seller is obliged to send the buyer notice of the consignment; but it is unclear whether this requirement is additional to délivrance or a part of it, and thus whether failure to send the notice prevents the passing of risk. On this point, the UNCITRAL Revision underwent a series of changes. The Working Group removed altogether the general section on unascertained goods-- an approach which was hardly satisfactory. Then UNCITRAL, at the final stage of drafting, reversed this policy and supplied a new provision, art. 79(2). When the goods are not identified to the contract by marking or otherwise, the passing of risk is delayed "until the seller sends the buyer a notice of consignment which specifies the goods." Dispatch of such a notice is required by the Draft Convention, and represents frequent commercial practice to ensure prompt delivery and keep down costs. But linking the passage of risk to the dispatch of notice was precisely the proposal rejected by the Working Group. One reason for their doing so was that it might lead to the transfer of risk after transit had begun. It is obviously difficult to determine the time when seawater entered the hold or seepage of the goods occurred in a long voyage, and it seems regrettable that the Draft Convention ignores these policy considerations.
A form of transaction which poses particular problems of risk is the sale of goods afloat. Consider the following example:
On April 15 S ships a cargo of sugar. On April 20 S sells the entire consignment to B. On arrival, the sugar is found to have been damaged by seawater.
It is undesirable for risk to be on S when the goods arrive, but for B to take the risk from April 20 would mean that risk passes mid-shipment; yet on April 15 there was no one to whom risk could be transferred. This subject led to long discussions among the draftsmen of ULIS. Their solution was to prescribe the retroactive passage of risk to the buyer as from the time when the goods were shipped. Although different from some other international formulations, this seems a commendable approach, and it has been preserved in art. 80 of the Draft Convention, with the sensible refinement that risk passes only from the handing over of the goods to the carrier who issues the documents. Where the seller knew or ought to have known that the goods had been lost or damaged and discloses this fact to the buyer, presumably the same retroactive transfer of risk occurs. But if there is no such disclosure, the seller retains the risk of "such loss or damage": art. 80 (2d sentence). This removes the impracticability of risk passing mid-shipment, which had here been the approach of ULIS. But the word such imports a difficult distinction between types of risk. In the example above, suppose S believed on April 20 that one tenth of the sugar had been damaged, but failed to disclose this to B. In fact, one fifth of the cargo had by then been affected. Presumably, the risk of the other tenth would rest on B. And if in the subsequent course of the voyage the dampness seeped through the rest of the sugar, would this damage fall to S or B? As it occurred after April 20, although from a cause known before the contract date, the wording of the Convention suggests it is the buyer's loss, which is surely unfortunate. If the sale to B was only of half the cargo, the goods would be unascertained, but the same retroactive assumption of risk would occur. A proposal to change the position where unidentified goods are concerned was rejected by the Committee "on the grounds that it would unnecessarily limit the scope of [art. 80]."  The result is in striking contrast with the position in pre-shipment sales. If S had already sold the sugar to B on April 10, risk would only pass on dispatch of the notice of consignment under art. 79(2).
Art. 81 of the Draft Convention is a residual provision, covering all circumstances outside arts. 79 and 80. It thus determines the passage of risk in a sale where there is no intervening carriage. In such a case, it seems clear that risk should be on the party in possession of the goods. He can take precautions to avoid hazards to the goods, and is better able to insure. The standard "building and contents" insurance policy will probably include the goods in the insured's possession, so he may have cover in any case, and for the other party to arrange insurance on the conclusion of the contract would probably be more expensive, as it would involve calculation of unknown risks (the kind of building in which the goods are stored, the nature of the other materials kept there, etc.). Moreover, if the other party suspects that the loss was not accidental but was caused by the possessor's negligence, this may be very hard to prove and could lead to costly litigation. This policy is realised by the UNCITRAL Draft Convention. Risk prima facie passes to the buyer "when the goods are taken over by him."  However, if the goods are stored in an independent warehouse or held by a bailee, delivery may be effected by endorsement of a warehouse receipt or by attornment. In that case it would be wrong to retain the risk on the seller until the buyer chooses to receive physical possession of the goods, which the phrase "taken over" suggests. The Convention therefore provides in art. 81(2) that risk then passes "when delivery is due and the buyer is aware of the fact that the goods are placed at his disposal. . . ." "Awareness" implies actual knowledge, so this appears to be an exception to the Draft Convention's general rule that failure to receive a notice properly sent by one party does not prevent him from relying upon it. If the goods are unascertained, they are deemed -- by art. 81(3) -- not to be placed at the buyer's disposal until they have been clearly identified to the contract.
Although drafted with a warehouse sale in mind, it appears that art. 81(2) will also apply to a sale involving carriage where the seller must deliver the goods at a particular destination -- a contract excluded from the general rules of arts. 79 and 80. Under ULIS, it is not clear whether risk passes only when the goods are actually handed over to the buyer at the destination, or whether placing the goods there at his disposal will suffice. By art. 81(2), the Draft Convention expressly provides the latter.
2. Breach by the SellerNon-conformity
Where the seller supplies goods which do not conform to the contract, ULIS -- as mentioned above -- subjects the transfer of risk to a special provision: art. 97(2). If the buyer declares the contract avoided or requires goods in replacement, the risk remains on the seller; if he does not, risk passes at the moment it would have passed had the goods conformed. Of all the provisions on risk in ULIS, this subsection has been the most criticized. Even Professor Tunc, the leading light in the final stages of drafting the Hague law, admits that it can give rise to difficulties. The passing of risk is made dependent on the buyer's remedies. In ULIS these are complex, and the remedy for non-performance of each type of obligation of the seller is separately stated: de minimis deviations are ignored, but much depends on the intricately defined concept of "fundamental breach." 
If the goods were produced by the seller, one remedy available to the buyer is to require the seller to repair the defects. This could operate as follows:
S is a manufacturer of polishing machines. In March, S sells 10 of his machines to B. On 5 April B receives the machines but they do not work properly. On 10 April B informs S of this, and on 20 April, after further consideration, instructs S to repair them. On 30 April, while S is in the process of curing the faults, the machines are seriously damaged by fire.
When B elects the remedy of repair, it seems that by art. 97(2) of ULIS the risk will be on him. After 30 April B may require a substitute delivery or avoid the contract, for the repair has not been, and now cannot be, completed. Under art. 97(2) the risk will then revert to S. But does it do so retroactively? Tunc says, with some hesitation, that it does not. This appears to be because the buyer, by initially requiring the seller to remedy the defects instead of to supply conforming goods, had "accepted" the machines tendered and so must bear the risk for the duration of his "acceptance." But this construction of art. 97(2) -- which makes no reference to acceptance -- is unclear indeed, and it is not surprising that other commentaries on the law give conflicting interpretations. The confusion is emphasized by the contrasting result if the fire occurred before 20 April, either during the short period within which B must notify S of the defects, or after B sent notice on 10 April but before he finally decides on the remedy of repair. B could then at once reject the machines. Tunc declares that under art. 97(2) the transfer of risk would then be negated -- so in this case it does revert retroactively! 
If the defect in the goods is latent, art. 97(2) again presents difficulties. ULIS expressly states that the buyer may rely on such a defect where the limitation period of two years has not expired and the buyer gives prompt notice upon the defect's discovery. If the buyer accordingly avoids the contract after a year, Tunc says that there has been no "valid acceptance of the goods and of the risk."  An alternative view is to regard the risk as passing to the buyer on the handing over of the goods, but reverting retroactively under art. 97(2) upon his subsequent avoidance. The second approach seems preferable, as after two years the buyer can no longer avoid and by then, obviously, the risk must be on him. Yet another solution is to hold that only the risks resulting from the defect revert to the seller. If the buyer had possession of the goods for six months, commercial practice might suggest that he should be responsible for such risks as fire and theft.
The confusion latent in ULIS art. 97(2) led UNCITRAL to abandon this approach. UNCITRAL's provisions determining the effect on risk of seller's breach underwent several redrafts. The flnal version, art. 82, merely declares that the other articles on risk do not impair the remedies available to the buyer upon a fundamental breach by the seller. Under the Draft Convention, any non-conformity constitutes a breach, but the buyer's right to reject the goods is restricted. Only if substantial damage is caused, amounting to "fundamental breach," can the buyer either avoid the contract or require the seller to perform his obligations by a substitute delivery. In order to rely on the non-conformity the buyer must give reasonable notice of the breach and, as in ULIS, a two-year limitation period applies. The system of remedies in the UNCITRAL revision is thus considerably simpler than in ULIS, and the definition of "fundamental breach" is markedly improved and more practical. In consequence, the effect of these remedies on risk is much easier to ascertain.
A frequent example is a shipment of non-conforming goods later damaged in transit. Under the Convention, if the defect amounts to a fundamental breach, the buyer on giving proper notice can avoid the contract or require conforming goods in replacement. In effect, the risk is therefore thrown back on the seller retroactively. If the breach is not fundamental, the buyer cannot reject the goods and the risk will pass to him on shipment in the usual way under art. 79(1). He can declare the price reduced and recover damages from the seller to account for the non-conformity  but not for the injury to the goods during transit. The policy behind this result seems the corollary of the buyer's restricted right to avoid the contract. The buyer cannot return the goods, so it is he who should cope with the loss, just as if conforming goods had been shipped.
The UNCITRAL revision by implication gives the buyer a right to ask the seller to cure the defects in the goods. The preceding example of a delivery of faulty machines may accordingly be considered again. If the breach is non-fundamental, asking S to repair is simply an alternative to demanding a reduction in the price; neither affects the passage of risk. If the breach is fundamental, B may not avoid the contract during the period allowed for performance but only within a reasonable time after that period has expired. Presumably, once the goods have been destroyed by fire, B may avoid the contract as performance cannot be completed. But does the risk, which passed to B on the handing over of the goods, revert retroactively to S? Or is B merely discharged from his obligation to pay the price while nonetheless being liable to S for the fire damage? This is the same problem as was found under ULIS. And there is similar uncertainty if the loss occurs during the short periods for examination or notice. It is disappointing to find that the Draft Convention has failed to remove one of the difficulties for which the analogous provisions of ULIS had been criticized.
Where the defect is latent, the UNCITRAL revision is a marked improvement on ULIS. Free from the shackles of délivrance, risk passes to the buyer in the normal way -- on the handing over of the goods. If the non-conformity is fundamental, then the buyer may rely upon it (as in ULIS) by giving reasonable notice after its discovery.  The risk is therefore thrown on the seller retroactively if and when the buyer avoids the contract. Latent and patent defects are accordingly treated alike.
Under ULIS, if the seller ships goods late, the facts do not fall within either paragraph of art. 97. The goods conform to the contract, so art. 97(2) is inapplicable, but the délivrance is not made "in accordance with the provisions of the contract," so art. 97(1) is not in point. The position is therefore not covered by the chapter on risk at all, and one must look to the buyer's general remedies. If the delay is a "fundamental breach," the buyer can avoid or require performance. Once the seller has delivered the goods (effectué la délivrance), this right to avoid is lost if not exercised promptly. If the breach is not fundamental, there is a Nachfrist provision by which the buyer can put the seller on notice: if no delivery is made within a specified grace period, the breach is converted into a fundamental one.
What then is the position regarding risk? Tunc thinks that if delay amounts to a fundamental breach, risk only passes when the buyer waives his right (renonce à son droit) to declare the contractavoided, and that it will only do so from that time (i.e., not retroactively). A specific example may assist:
(i) The contract is for a shipment sale, specifying that the goods should arrive in the first week in April. S ships the goods late and they only arrive on 11 April. On 12 April, before B can reject the goods, they are destroyed by fire.
If this amounts to a fundamental breach, the risk will be on S, although B has received conforming goods into his possession. And this will still be the case if the goods are only slightly damaged and B chooses to keep them! He can recover from S for the fire damage as he still has the right to avoid. The unfortunate nature of such a rule emerges if the facts are slightly changed:
(ii) On 9 April B contacts S to ask why he has not received the goods. He is told that they were shipped late and will arrive on 12 April. B agrees to accept this shipment. The goods only arrive on 14 April. On 15 April, before B can reject the goods, they are destroyed by fire.
Presumably, when B agrees on 9 April to take the shipment, this amounts, within Tunc's terms, to a waiver of his right to avoid. If the risk passes then, it does so in mid-transit. But as the goods do not arrive until 14 April, B may avoid after all, so has there in fact been such a waiver?
If the breach is not fundamental, Tunc states decisively that risk passes at the time of delivery. But this result is not explicitly stated by ULIS and also creates problems. Suppose that in example (ii), B on 9 April effectively grants S a four day grace period. By Nachfrist, a fundamental breach is created on 12 April when the goods fail to arrive and B may then avoid the contract. If the risk passed to B on shipment, when does it return to S -- on 11 April or on B's subsequent declaration of avoidance? Does it do so retroactively and if so, to what date?
The contrast displayed by the Draft Convention is striking. For a time the Working Group considered a separate provision governing risk by delay, but the final draft embraces such breach in the single art. 82 discussed above. The transfer of risk does not have to be elusively inferred from the buyer's remedies; instead, those remedies impinge upon the risk, which is readily located under the rules of arts. 78-81. The consolidation of all the remedies of the buyer upon seller's breach is one of the major advances of the UNCITRAL revision. His remedies for delay are accordingly the same as those for non-conformity, with one significant addition. Even if the delay does not constitute a fundamental breach, the buyer may require performance, and if he fails to receive it within the specified time -- or if the seller refuses his request -- he can avoid the contract. Thus the Nachfrist provisions of ULIS, abandoned in the case of non-conformity, is preserved by the Draft Convention in the event of delay.
The two examples above may be considered again. In both, the risk passes to B on shipment by art. 79(1). In (i), if the three-day delay amounts to a fundamental breach, B has a right to avoid. One day seems a reasonable time for the exercise of this right, the destruction of the goods presents no obstacle, so by B's avoidance risk reverts retroactively to S. In (ii), if the breach is non-fundamental, the agreement to accept the goods on 12 April constitutes an additional time for performance. The failure of the goods to arrive within that time is a breach entitling B to avoid the contract, but it is not a fundamental breach under the terms of the Draft Convention. Does B's avoidance therefore throw the risk back upon S? One would suppose it does, in the usual way, except for the troublesome wording of art. 82. This states that "[i]f the seller has committed a fundamental breach of contract . . ." the other provisions on risk do not interfere with the buyer's remedies. So it could be interpreted to mean that on a non-fundamental breach, the fact that the risk has prima facie passed to B under art. 79(1) does impair his right to declare the contract avoided.
The obvious solution is to delete the word "fundamental" from art. 82. And indeed it appears that precisely this was UNCITRAL's intention! The report of the final stage of drafting clearly indicates that such a proposal was adopted, but in the subsequently published texts, "fundamental" has been retained. Presumably this is no more than an oversight, which one hopes will soon be corrected. If it is, then in every case of delay the Draft Convention will place the entire risk on the buyer if the goods are ultimately accepted, and on the seller if they are rejected -- an eminently sensible result.
Both ULIS and the Draft Convention include provisions concerning the seller's right to deliver the goods before the contract date. There seems no reason why such a right should confuse the question of risk, but under the peculiar phraseology of ULIS art. 97(1) it will, for early delivery is not "delivery . . . in accordance with the provisions of the contract." Whether risk passes upon such delivery is accordingly uncertain. By contrast, under the UNCITRAL revision, there is clearly nothing to interfere with the transfer of risk in the normal way under arts. 79 or 81(1). However, a problem arises in the case of a "destination sale" which does not come within the scope of art. 79 and therefore falls within art. 81(2).[75a] Under that provision, risk passes "when delivery is due and the buyer is aware of the fact that the goods are placed at his disposal. . . ." The due delivery requirement seems to mean that if early delivery is made and the buyer chooses not to take possession of the goods, he can keep the risk on the seller -- a result which is perhaps only reasonable. However, if the buyer does choose to take the goods, the wording of the article appears to keep the risk on the seller nonetheless! It appears doubtful that this surprising consequence was intended by the draftsmen.
The most common form of buyer's breach in a sales contract concerns payment, and raises no question of risk. But the buyer is also under an obligation to receive delivery of the goods, and if he delays, problems of risk arise. To resolve them simply poses the difficulty of embracing different kinds of contract and differing types of goods within a single article.
Under ULIS, where the buyer causes delay in the handing over of the goods, there is by definition no délivrance, for the goods have not been handed over. So the uniform law deals with this situation in a separate provision, art. 98. Where a contract for the sale of specific goods does not call for shipment, if the buyer's delay constitutes a breach, risk passes to him "from the last date when, apart from such breach, the handing over could have been made in accordance with the contract."  This seems a wise approach, as it prevents the buyer profiting from his breach by keeping the risk on the seller, notwithstanding that the risk is on the buyer while the seller has possession of the goods. In a sale of unascertained goods, ULIS expressly imposes two requirements for the passage of risk. The goods must be "manifestly appropriated to the contract" (which requires an overt act) and, in addition, risk can only pass when the buyer has been notified of such appropriation. The notification serves as a warning to the buyer that the risk is on him, and prevents a fraudulent seller claiming that goods destroyed had already been appropriated to the contract, which it would be very hard for the buyer to disprove. This requirement, the Special Commission reporting on the 1956 Draft remarked, is "an innovation not known to the majority of legal systems." 
The UNCITRAL revision distinguishes between contracts where the buyer is required to receive the goods at the seller's place of business and those where he is to receive them elsewhere. For the first case, art. 81(1) preserves the ULIS solution. For the second case, art. 81(2) prescribes that risk passes when delivery is due and the buyer is aware of the fact that the goods are at his disposal. A notification requirement is thus introduced which covers certain sales of specific goods. More significantly, if the contract allows a period for delivery, risk will pass at the end of that time if the goods are kept on the seller's premises, but may be transferred at the start of it if they are not.
Regarding unascertained goods, the Draft Convention makes several changes. Tbere was criticism of the ULIS terminology, since "unascertained" and "appropriation" are terms which have complex connotations in several national laws; hence the expression "identification" has been used instead. Secondly, although the goods are deemed not to be at the buyer's disposal until clearly identified to the contract, when the buyer is to receive delivery at the seller's place of business the notification requirement has been removed. It is hard to see the rationale of this shift in policy, and its development has been tortuous. The earlier drafts of the Convention included a general notification requirement, which was deleted altogether at the last session of the UNCITRAL Working Group, only to be partially reintroduced into the final draft by the Committee of the full Commission.
Once again, an illustration may help to show the effect of these provisions:
The contract is for a sale of 200 tons of potatoes to be delivered the third week in April. S identifies goods to the contract by storing the potatoes in labelled sacks on 14 April. B fails to collect them by 30 Apri1 when the warehouse in which they are stored is destroyed by fire.
Under ULIS, risk passed to B on 21 Apri1 if S had notified him of the identification. Under the UNCITRAL draft, notification is irrelevant if the warehouse belongs to S; but if it is a public warehouse, S would have to notify B that the goods were ready for collection, and by doing so could already pass the risk to B on 14 April -- a distinction which would assume still greater significance if the fire were on 19 April!
Suppose the sale was not of potatoes but of oil, part of a larger quantity S stored in his tanks. Identification of the goods to the contract in those circumstances is not possible. ULIS provides for such a case in art. 98(3): to pass the risk it is sufficient for the seller to do "all acts necessary to enable the buyer to take delivery."  Thus the risk would pass to B on 21 April nonetheless. The UNCITRAL Draft Convention contains no similar provision; the oil is accordingly deemed not to be placed at B's disposal, and the risk remains on S.
However, the most curious aspect of risk on buyer's breach in the Draft Convention concerns a shipment contract. When the seller is required to ship the goods and the buyer is required to name a vessel for the purpose (as in many FOB sales), the contract falls within art. 79, and it seems clear that it is not governed by the provisions of art. 81(2). In consequence, if the seller is unable to send the goods because the buyer fails to appoint a ship, the risk remains on the seller -- and where the goods are perishable this includes the risk of deterioration which results directly from the buyer's breach. To relegate the seller to his remedy in damages seems an unfortunate approach, and differs from the position in most municipal systems.
The handling of risk in the Draft Convention can thus be criticized, particularly with regard to a buyer's breach of contract. Nonetheless, UNCITRAL's achievement in dealing with risk is notable. The general advance on ULIS should be evident, and it must be remembered that the Convention is not yet in final form; further amendments may be proposed at the forthcoming diplomatic conference. Moreover, the provisions discussed above will not apply to all sales covered by the law. The Draft Convention, like ULIS, sets out non-mandatory rules; they may be displaced by a contrary intention of the parties. Where risk is concerned, such exclusion seems especially likely. The Draft Convention follows ULIS in declining to define any of the trade terms in frequent commercial use. In an FOB or CIF sale, if the contract is ex ship or ex works, or where any of the other standard formulations is employed, the meaning of these terms will determine the transfer of risk. Indeed, the Draft Convention contains an article expressly binding the parties to a commercial usage on which they have agreed.
The adoption of this approach by ULIS provoked much academic discussion. It was argued that the absence of specific definitions deprived the chapter on risk of much value and rendered the law an incomplete regulation of international sales. The parties will often insert such a term, giving rise to the very uncertainty of interpretation that a uniform law was intended to avoid. This is especially the case in a law designed to apply in developing countries, where few lawyers may be specialists and even merchants may be uninformed of trading practices well-known among the developed nations.
However, not only can definitions of trade terms become very elaborate, but their meaning is subject to development, whereas an international Convention cannot readily be revised to keep pace with changing commercial customs. Two of the leading current definitions, the ICC's Incoterms (1953) and the Revised American Foreign Trade Definitions (1941) differ in certain respects, so there does not appear to be a generally accepted meaning of all the terms. And there is sometimes uncertainty over what merchants intend by these expressions; they may use them very differently from lawyers, and placing the formulations in a legal straightjacket may therefore not assist in ascertaining the parties' intention. It is one thing to define the terms which the draftsmen use themselves, but quite another, and more hazardous, task to supply definitions for terms employed by others. Significantly, none of the Governments in the comments submitted to UNCITRAL on ULIS complained of the lack of definitions of commercial terms.
The function of a uniform law which prescribes a statutory norm is not the same as that of standard definitions which the parties may voluntarily adopt. And a law designed for world-wide use requires greater flexibility than a merely regional formulation. It is from the combination of a global uniform law, regional measures and commercial practices that what Schmitthoff has called the new law merchant derives its character and makes its contribution. There may be scope for UNCITRAL to assume a co-ordinating role in the preparation of a body of trading terms of general acceptance. But that is a task distinct from the drafting of a uniform law of sale. The Convention on Contracts for the International Sale of Goods will serve to provide a general referent in case of need, and establish, in the words of one of its architects, ". . . a framework on which a common body of international doctrine and case-law can be built." 
1. Diplomatic Conference on the Unification of Law Governing the International Sale of Goods, The Hague, 2-25 April (1964), II Documents 43 (1966): "Report of the Special Commission 1956," para. 6.
2. See Atiyah, The Sale of Goods 164 (5th ed. 1975).
3. The difficulty of treating risk as a "lump-concept" has been pointed out in the context of ULIS: Berman, "The Uniform Law on the International Sale of Goods: A Constructive Critique," 30 Law & C.P. 354, 364 (1965).
4. Corresponding statutes have been enacted in India, New Zealand and Australia.
5. For a concise, but dated, survey see Note, "Risk of Loss in Sales," 6 Tul. L. Rev. 272 (1932).
6. See Takinawa, "Risk of Loss in Japanese Sales Transactions," 42 Wash. L. Rev. 463 (1967).
7. This view of sales law is canvassed by Honnold, "The Uniform Law for the International Sale of Goods: the Hague Convention of 1964," 30 Law & C.P. 326, 338 (1965).
8. Farnsworth, "Unification of Sales Law at the Regional and International Level: Why They Behave Like Americans," in Aspects of Comparative Commercial Law: Sales, Consumer Credit and Secured Transactions (Ziegel & Foster eds.) 110, 120 (1969).
9. The West German BGB has been criticized as still further divorced from commercial needs than English law, thus requiring extensive "construction" by the courts: Zweigert, "Aspects of the German Law Relating to Sale of Goods," in Some Comparative Aspects of the Law Relating to the Sale of Goods, Int. & Comp. L.Q., Supp. Pub. No. 9 (1964) 1, 14.
10. "The Hague Conference on the Unification of Sale Law," 1 Am. J. Comp. L. 58, 61 (1952).
11. Rabel, 3 Conflict of Laws 92 (2d ed. 1964).
12. ULIS art. 8; Draft Convention art. 6(b).
13. The Convention will apply to some consumer sales: art. 2(a) provides that goods bought for personal, family or household use are covered if the seller neither knew nor ought to have known of this fact. Many non-mercantile purchases (e.g., a sale of soap to a business for the personal use of its employees) will be within the law. See also art. 1(3).
14. ULIS added the qualification "or of some other person for whose conduct the seller is responsible" (art. 96). This phrase has been removed as superfluous; it is a general principle of the law that a party’s liability for his acts extends to the acts of those for whom he is responsible.
15. Tunc, Commentary on the Hague Conventions of 1st July 1964 on the International Sale of Goods and on the Formation of Contracts of Sale 46 (1964) [hereinafter "Tunc"].
16. Arts. 71-2.
17. For a sustained criticism of the concept of déliverance, see Honnold, "A Uniform Law of International Sales," 107 U. Pa. L. Rev. 299, 318, 325-6 (1959); Report of the Secretary-General, "Delivery in the Uniform Law on the International Sale of Goods (ULIS)" (A/CN.9/WG.2/WP.8) paras. 9-18, III Yearbook 3 (1972).
18. Art. 97(2).
19. Report of the Secretary-General, "Analysis of replies and comments by governments on the Hague Conventions of 1964" (A/CN.9/31) paras. 98-104, I Yearbook 170-1; Note by the Secretary-General, "Analysis of comments and proposals relating to articles 18-55 of ULIS" (A/CN.9/WG.2/WP.10) paras. 9-18, III Yearbook 54, 56-8.
20. Strictly, UNCITRAL has no authority to revise ULIS, and the Working Group appointed by the Commission at its second session was technically instructed to consider ULIS and "ways and means by which a more widely acceptable text might be prepared and promoted. . . ." UNCITRAL, Report on Second Session (1969) A/7618, para. 38, I Yearbook 100. Examining ULIS article by article, the Working Group concluded that a new law was required, and as the process of drafting inevitably used ULIS as a foundation, it is convenient to speak of a "revision."
21. Working Group, Report on Third Session (1972), Annex II, para. 17 (A/CN.9/62, Add. 1 & Add. 2), III Yearbook 77, 83-4.
22. In ULIS the outcome is determined by a combination of arts. 97(1) and 19(2).
23. This was not stated in ULIS; it was added at the final stage of drafting the Convention, at the 10th Session of UNCITRAL; "Report of Committee of the Whole I relating to the Draft Convention on the International Sale of Goods," paras. 538-9, UNCITRAL, Report on Tenth Session (1977) Annex I (A/32/17) 146.
24. Id. paras. 535-6.
25. ULIS leaves this uncertain, as the Austrian observations to the 1964 Hague Conference pointed out. The ICC there submitted an amendment to supply the deficiency but it was not adopted: Ellwood, "The Hague Uniform Laws governing the International Sale of Goods," in Some Comparative Aspects of the Law Relating to the Sale of Goods, supra n. 9 at 47.
26. Art. 19(3). The commentaries on the law differ: Tunc 47-8 (the English version is in part a mistranslation) declares it is not part of the obligation to "deliver"; similarly, Mertens & Rehbinder, Internationales Kaufrecht, Kommentar zu den Einheitlichen Kaufgesetzen 150 para. 14; 299 para. 6 (1975); the notice requirement thus has no effect on the transfer of risk. Per contra Graveson, Cohn, Graveson, The Uniform Laws on International Sales Act 1967, 64 (1968). Dölle (ed.), Kommentar zum Einheitlichen Kaufrecht 181-3, paras. 104-19 (1976), considers that while not practically a constituent of delivery, it is a condicio iuris of delivery, and that risk passes upon the sending of the notice retroactively as from the time of the handing over of the goods.
27. The Commission at its 10th Session established a Committee of the Whole I to consider the draft prepared over the previous 7 years by the Working Group; the Committee’s report and recommendations were adopted by the Commission: UNCITRAL, Report on Tenth Session (1977).
28. Art. 30(1).
29. Working Group, Report on Fifth Session (1974) A/CN.9/87, para. 234, V Yearbook 29, 50.
30. Tunc, 106.
31. Art. 99(1); cf. UCC 2-509(1)(a), where it is unclear whether or not risk passes retroactively, although Comment 2 to the section declares unequivocally that it does not.
32. E.g., cf. ECE Conditions No.1A, para. 9.2: "Where the contract is concluded only after the time when the goods actually pass the ship’s rail at the port of shipment, the risk shall pass from the seller to buyer on conclusion of the contract."
33. Draft Convention art. 80: "The risk in respect of goods sold in transit is assumed by the buyer from the time the goods were handed over to the carrier who issued the documents controlling their disposition. However, if at the time of the conclusion of the contract the seller knew or ought to have known that the goods had been lost or damaged and he has not disclosed such fact to the buyer, such loss or damage is at the risk of the seller."
34. This was added by the Committee of the Whole; the version proposed by the Working Group would have retained the entire risk on the seller. Report of Committee of the Whole I, supra n.23, paras. 534-46 at 145-7.
35. Id. para. 540.
36. See Report of the Secretary-General, "Issues Presented by Chapters IV-VI of ULIS" (A/CN.9/87, annex IV) paras. 70-5, V Yearbook 91.
37. Art. 81(1). ULIS achieves the same result by a combination of arts. 97(1) and 19(1).
38. The Draft Convention allows for this in art. 29(b): "Placing the goods at buyer’s disposal."
39. Art. 10.
40. Report of Committee of the Whole I, supra n. 23, paras. 550-2 at pages 148-9.
41. Tunc, 102.
42. Art. 10. For strident criticism Graveson, Cohn, Graveson, supra n. 26 at 55-7; Sutton, "The Hague Convention of 1964 and The Unification of the Law of International Sale of Goods," 7 U. Qld. L.J. 145, 168-9 (1971).
43. Art. 42(1)(a).
44. Arts. 41-4.
45. Tunc, 102 ("it would seem that it must be agreed, although the Article does not expressly say so . . .").
46. Graveson, Cohn, Graveson, supra n. 26 at 108 follows Tunc; contra: Dölle (ed.), supra n. 26 at 627, para 46. Mertens & Rehbinder, supra n. 26 at 301 paras. 18-9, argue that risk never passed to the buyer at all because a demand for repair is akin to a request for replacement delivery.
47. Art. 39.
48. Arts. 42(2), 43, 44(2).
49. Tunc, 102-4; similarly, Dölle (ed.), supra n. 26 at 627, paras. 48-9, Mertens & Rehbinder, supra n. 26 at 300, paras. 12-3.
50. Art. 39(1).
51. Tunc, 104.
52. But see infra, text at n. 73.
53. Arts. 41-45; in certain circumstances the seller has a right to make a substitute delivery, without this being requested.
54. Arts. 37-38.
55. Art. 23. See the article by Michida, II-E supra.
56. Arts. 27-8, 31(1)(a).
57. The buyer nonetheless has specific obligations to salvage the goods: arts. 74-77.
58. Arts. 41, 46. On reduction of the price see the article by Bergsten & Miller, II-D supra.
59. Art. 42(1) is intended to embrace a request for repair, though this construction is less than clear: Report of Committee of the Whole I, supra n. 23, paras. 243-58 at p. 89-92.
60. Arts. 43(2), 45.
61. Art. 37(1); destruction of the goods is no bar to avoidance: art. 67(2).
62. Art. 26(1). A supplementary definition of fundamental breach is supplied: art. 28.
63. Art. 26(3).
64. Art. 27.
65. Tunc, 100.
66. Id.; cf. Graveson, Cohn, Graveson, supra n. 26 at 108, which seems to suggest the contrary.
67. See Revised text of the Convention on the international sale of goods as approved or deferred for further consideration by the Working Group on the International Sale of Goods at its first six sessions (A/CN.9/100, annex I) art. 69(2), VI Yearbook 69.
68. Arts. 43, 45(1)(b).
69. ULIS arts. 27, 44(2); cf. Draft Convention arts. 42(2), 45(1).
70. Art. 45(1)(a).
71. Arts. 45(2)(a), 67(2)(a), 82.
72. ULIS art. 29, Draft Convention art. 48(1).
73. UNCITRAL, Report on Tenth Session (1977) Annex I (A/32/17) 149 para. 555 [the present art. 82 was then numbered 68]; this view seems confirmed by the summary record: UNCITRAL, Tenth Session, Committee of the Whole I, Summary Record of the 29th Meeting, 10th June 1977 (A/CN.9(X)/C.1/SR.29) paras. 20-28. But cf. Text of the Draft Convention on the International Sale of Goods, UNCITRAL, Report on Tenth Session (1977) (A/32/17) 26; Text of The Draft Convention on Contracts for the International Sale of Goods, UNCITRAL, Report on Eleventh Session (1978) (A/33/17) 30.
74. ULIS art. 29, Draft Convention art. 48(1).
75. Tunc, 100, says risk is only transferred when the due date for delivery arrives. Graveson, Cohn, Graveson, supra n. 26 at 108 declares it passes on the buyer’s acceptance of the goods; so also Mertens & Rehbinder, supra n. 26 at 302, para. 21.
75a. See text supra at n. 40.
76. Art. 48(1).
77. A slight amendment to art. 81(2) of the Draft Convention would cure this: ". . . the risk passes when the buyer takes over the goods, or when delivery is due and the goods are placed at his disposal at that place, whichever is the earlier."
78. ULIS art. 98(1).
79. Art. 98(2); cf. the ambiguity of ULIS in the case of seller’s breach: see discussion at n. 26.
80. Report of the Special Commission, supra n. 1 at 43-4 on art. 110.
81. See art. 31(b).
82. Report of the Secretary-General, supra n. 36 at para. 84, V Yearbook 92.
83. Art. 81(3).
84. The unfortunately brief report does not explain why: Working Group, Report on Seventh Session (1976) A/CN.9/116, VII Yearbook 96.
85. It appears that here notification is not required, which seems surprising.
86. The position of art. 81(2), its wording, and the drafting history all confirm this view: Report of the Secretary-General, supra n. 36 at para. 82; Working Group, Report on Fifth Session, supra n. 29 at paras. 226, 233; Report of Committee of the Whole I, supra n. 23 at paras. 550-2.
87. E.g., Sale of Goods Act 1893 s.20, BGB § 324; cf. UCC 2-510(3) -- the risk of deterioration is unlikely to be covered by the seller’s insurance.
88. ULIS art. 3; Draft Convention art. 5.
89. Art. 8; cf. ULIS art. 9. For a discussion of the problems of construction, see Jokela, "The Role of Usages in the Uniform Law on International Sales," 10 Scand. Stud. L. 81 (1966).
90. See, e.g., IALS Colloquium, Summary of Proceedings, Unification of the Law Governing International Sale of Goods (Honnold ed.) 380-6 (1966).
91. Both the Comecon (CMEA) and UN ECE Conditions have detailed provisions particularizing the requirements for different kinds of shipment.
92. See, e.g., Comptoir d’Achat et de Vente de Boerenbond Belge S.A. v. Luis de Ridder Limitada (The Julia),  A.C. 293, where in a sale stated to be "CIF Antwerp," the House of Lords held that, looking at the substance of the transaction, this was not a true CIF contract.
93. "Analysis of replies and comments by government . . .," supra n. 19.
94. "The Unification of the Law of International Trade" (Travers Lecture), (1968) J. Bus. L. 105, 111-2.
95. Honnold, "A Comparison of National and Regional Unifications of the Law of Sales and Avenues Towards Their Harmonization: Prospects and Problems," in Unification of the Law Governing International Sale of Goods, supra n. 90 at 23.
Go to Database Directory || Go to CISG Table of Contents