Section C. Quarter under unusual circumstances of combat
Additional Protocol I
Article 41(3) of the 1977 Additional Protocol I provides:
When persons entitled to protection as prisoners of war have fallen into the power of an adverse Party under unusual conditions of combat which prevent their evacuation … they shall be released and all feasible precautions shall be taken to ensure their safety.
Article 60 of the 1863 Lieber Code provides: “[A] commander is permitted to direct his troops to give no quarter, in great straits, when his own salvation makes it impossible
to cumber himself with prisoners.”
(emphasis in original)
Bosnia and Herzegovina
The Instructions to the Muslim Fighter (1993) issued by the Army of the Republic of Bosnia and Herzegovina in 1993 state that it is “left to the military command’s discretion to decide whether it is more useful or in the general interest to free, exchange or liquidate enemy prisoners of war”.
Burundi’s Regulations on International Humanitarian Law (2007) states:
If the conditions do not allow for the evacuation of prisoners of war, they must … be guarded while waiting until their evacuation is possible … [or] they must be released, while taking certain precautions for the security of [both the capturing forces] and of the prisoners of war.
The Regulations also states that “[i]t is prohibited to attack … the wounded, sick, shipwrecked, [or] prisoners of war”.
Canada’s LOAC Manual (1999) states:
Where persons entitled to protection as prisoners of war (PWs) have fallen into the power of an adverse party under unusual conditions of combat that prevent their evacuation as provided for in [the 1949 Geneva Convention III], they shall be released and all feasible precautions shall be taken to ensure their safety.
The “unusual conditions of combat” may include, for example, the capture of a PW by a long-range patrol that does not have the ability to properly evacuate the PW. In such circumstances, there would be an obligation to release the PW and take all feasible precautions to ensure his safety. Such precautions might include providing the PW with sufficient food and water or other aids to assist in rejoining unit lines.
Canada’s LOAC Manual (2001) states in its chapter on land warfare:
1. Where persons entitled to protection as prisoners of war (PWs) have fallen into the power of an adverse party under unusual conditions of combat that prevent their evacuation as provided for in the GIII, they shall be released and all feasible precautions shall be taken to ensure their safety. For the obligations of a belligerent with respect to the evacuation of PWs, refer to Chapter 10 (Treatment of Prisoners of War).
2. The “unusual conditions of combat” may include, for example, the capture of a PW by a long-range patrol that does not have the ability to evacuate the PW properly. In such circumstances, there would be an obligation to release the PW and take all feasible precautions to ensure his safety. Such precautions might include providing the PW with sufficient food and water or other aids to assist in rejoining unit lines.
France’s LOAC Manual (2001) states: “When the capturing unit is not able to evacuate its prisoners or to keep them until the evacuation is possible, the unit must free them while guaranteeing its own and the prisoners’ security.”
Ireland’s Basic LOAC Guide (2005) states:
Troops holding PWs [prisoners of war] may not kill them to preserve themselves, even in cases of extreme necessity; for instance, because they slow up their movements, require guards needed for other purposes, consume scarce supplies or appear certain to be set free by their own forces. If they cannot be held, they must be released unharmed.
Israel’s Manual on the Laws of War (1998) states:
Considerations such as the delay involved in guarding prisoners of war as opposed to the attainment of an objective, or even the allocation of manpower for transferring them to the rear line, do not permit the harming of prisoners who surrendered and were disarmed. It is hard to imagine a military mission so urgent as to render impossible the evacuation of prisoners to the rear or even binding them until additional forces arrive and which justifies their murder.
Israel’s Manual on the Rules of Warfare (2006) states:
Considerations such as delays in reaching the objective due to the need to guard prisoners-of-war or even assigning personnel for the purpose of transferring them to the home front, do not permit attacking prisoners-of-war who have surrendered and have been disarmed. It is difficult to think of a military mission of such urgency that it does not allow the removal of prisoners-of-war to the rear or even immobilising them until the arrival of reinforcements.
The Manual on the Rules of Warfare (2006) is a second edition of the Manual on the Laws of War (1998).
Kenya’s LOAC Manual (1997) states that, when the capturing unit, such as a small patrol operating in isolation, is not in a position to evacuate prisoners, “that unit shall release them and take precautions: a) for its own safety …; and b) for the released’s safety (e.g. giving them water and food, the means to signal their location, and subsequently providing information about their location to rescue teams)”.
Mexico’s Army and Air Force Manual (2009), in a section on the 1949 Geneva Convention III, states:
The general principles protecting prisoners of war include the following:
A. Prisoners of war must not be unnecessarily exposed to danger while awaiting evacuation from a fighting zone. When they have been captured in unusual circumstances, which prevent them from being evacuated in the normal way, they must be released and all feasible precautions taken to ensure their safety.
The Military Manual (1993) of the Netherlands provides that, when a person falls into the hands of the adversary under exceptional circumstances preventing his evacuation as a prisoner of war, this person must be released. This situation can occur, for instance, for a long-range post.
The Military Manual (2005) of the Netherlands states: “If a person falls into the adversary’s hands and the conditions of battle prevent that person from being removed as a prisoner of war, that person must be released.”
Sierra Leone’s Instructor Manual (2007) states:
In any combat operation, when enemy combatants are first captured they straightaway become POWs [prisoners of war]. The first action to be taken therefore is to:
f. Evacuate them as soon as the tactical situation permits. If you have to abandon them, you can tie them only
if left where they can be found; or release them and give them food and water.
[emphasis in original]
Spain’s LOAC Manual (1996) states that when the conditions of combat make it impossible to treat prisoners of war properly and to evacuate them (e.g. isolated special operations, small units, mass capture which exceeds the possibility of the unit in question), the prisoners must be released and all feasible precautions must be taken to ensure their safety.
Spain’s LOAC Manual (2007) states:
When the conditions in which the hostilities are being conducted prevent prisoners of war from being treated as required and evacuated (isolated action by special operation forces or small units, the capture of large numbers of prisoners of war that the detaining unit does not have the capacity to handle, etc.), they must be released and all feasible precautions taken to ensure their safety.
Switzerland’s Basic Military Manual (1987) provides: “If a commando raids an enemy post and captures soldiers by surprise without being able to take them along with it in its retreat, it shall not have the right to kill or injure them. It may disarm them, but it shall free them.”
United Kingdom of Great Britain and Northern Ireland
The UK Military Manual (1958) provides:
A commander may not put his prisoners of war to death because their presence retards his movements or diminishes his power of resistance by necessitating a large guard, or by reason of their consuming supplies, or because it appears that they will regain their liberty through the impending success of the forces to which they belong. It is unlawful for a commander to kill prisoners of war on grounds of self-preservation. This principle admits of no exception, even in the case of airborne or so-called commando operations …
Whether a commander may release prisoners of war in the circumstances stated in the text is not clear … If such a release be made, it would seem clear that the commander should supply the prisoners with that modicum of food, water and weapons as would give them a chance of survival.
United Kingdom of Great Britain and Northern Ireland
The UK LOAC Manual (2004) states:
“When persons entitled to protection as prisoners of war have fallen into the power of an adverse Party under unusual conditions of combat which prevent their evacuation as provided for in Part III, Section I, of the Third Convention, they shall be released and all feasible precautions shall be taken to ensure their safety.” … This rule covers cases such as commando operations or long-range reconnaissance patrols in enemy held territory.
In its chapter on air operations, the manual states:
12.64. Although it is forbidden to “kill or wound an enemy who, having laid down his arms, or having no longer means of defence, has surrendered at discretion”, in air-to-air combat, surrender is usually impracticable and occurs very infrequently.
12.64.1. In the special circumstances of air-to-air combat the continuation of an attack after an indication by the opponent of a wish to surrender is not inconsistent with the rule in paragraph 12.64 as the enemy pilot who remains in his aircraft cannot be said to have “laid down his arms” or to have “no longer a means of defence”. However, if the surrender is offered in good faith and in circumstances that do not prevent enforcement, for example, when the engagement has not taken place over enemy territory, it must be respected and accepted. Surrenders of enemy aircraft and crews should not be discouraged because not only is a psychological advantage gained, but an enemy aircraft and defecting aircrew can provide intelligence which, if promptly and properly evaluated, may be of inestimable benefit to operations planning.
Aircraft in distress
12.66. If it is clear that an enemy aircraft is disabled beyond recovery and is unable to continue and will not resume combat, or indeed is unarmed, and its immediate destruction would offer no military advantage then the attack must be broken off to allow its crew and passengers to evacuate.
12.66.1. A topic very closely related to that of surrender in the air is whether an attack should be continued against a disabled aircraft. The rule of the law of armed conflict forbidding the killing or wounding of an enemy who is hors de combat
is difficult to apply in the context of aerial warfare. Sometimes pilots simulate disability or loss of control. It is frequently necessary in aerial combat to pursue to destruction an enemy aircraft apparently in distress because of the impossibility of verifying its condition. Despite its apparent condition the aircraft may not have lost the use of all its offensive systems.
With regard to “downed” airmen in enemy-held territory, the manual states:
12.68. On land, a “downed” airman from an aircraft in distress must be afforded a reasonable opportunity to surrender before the attack upon him is resumed. Clearly if the “downed” airman is incapacitated he is hors de combat and the general rule will apply. The attack may be resumed immediately if he offers violence, attempts to escape or if, suffering no incapacity, he is in territory controlled by his own forces. The pilot who has crash-landed his aircraft and is attempting to complete its destruction or the destruction of any part of his or its equipment is committing a hostile act and may be attacked immediately.
Combat rescue of downed aircrew on land
12.69. The use of, for example, military assets to rescue aircrew who have been “downed” on territory under the control of the enemy is a combat activity. It is therefore legitimate for an enemy in such circumstances to attack the rescuers or by some other means to impede or prevent the rescue activity. However, that would not apply in the case of medical personnel, units or transports collecting the sick or wounded, see chapter 7.
12.69.1. The mere fact that a rescue service saves enemy personnel as well as its own does not entitle it to immunity from attack. Once taken prisoner, such enemy personnel should be accorded all the rights of a prisoner of war, at least until their status has been determined. Arguably, any communication by downed aircrew with their national authorities constitutes a hostile act, which would at that point justify an attack upon that aircrew. Such justification may also arise from other conduct indicating that the downed individual is seeking to continue the fight.
Combat rescue at sea
12.70. A “downed” airm[a]n at sea falls within the definition of “shipwrecked” so that he must not be attacked if he refrains from hostile acts.
12.70.1. Although it can be argued that search and rescue by aircraft used exclusively for rescuing airmen “downed” at sea in areas either not controlled by friendly forces or not occupied by enemy forces should be regarded as a protected activity under the provisions of Geneva Convention II 1949, it has been the general practice not to afford such protection.
In its chapter on prisoners of war, the manual states: “It is unlawful
to kill prisoners of war on grounds of self-preservation or because holding them would impede or endanger military operations.”
(emphasis in original)
United States of America
The US Field Manual (1956) states:
A commander may not put his prisoners to death because their presence retards his movements or diminishes his power of resistance by necessitating a large guard, or by reason of their consuming supplies, or because it appears certain that they will regain their liberty through the impending success of their forces. It is likewise unlawful for a commander to kill prisoners on grounds of self-preservation, even in the case of airborne or commando operations.
Denmark’s Military Criminal Code (1973), as amended in 1978, provides:
Any person who uses war instruments or procedures the application of which violates an international agreement entered into by Denmark or the general rules of international law, shall be liable to the same penalty [i.e. a fine, lenient imprisonment or up to 12 years’ imprisonment].
Denmark’s Military Criminal Code (2005) provides:
Any person who deliberately uses war means [“krigsmiddel”] or procedures the application of which violates an international agreement entered into by Denmark or international customary law, shall be liable to the same penalty [i.e. imprisonment up to life imprisonment].
Under Ireland’s Geneva Conventions Act (1962), as amended in 1998, any “minor breach” of the 1977 Additional Protocol I, including violations of Article 41(3), is a punishable offence.
Japan’s Law concerning the Treatment of Prisoners of War and Other Detainees in Armed Attack Situations (2004) states:
When there are reasonable grounds to believe that the captive persons cannot endure the transfer for delivery in the light of the physical and mental condition of the captive persons, available means of transport and other circumstances, the SDF [Self-Defense Force] personnel … may … release the captive persons immediately, after taking all feasible precautions in accordance with the relevant situation, to ensure their safety, such as bringing them to nearby places where they can avoid the direct danger from combat actions, provision of appropriate medical supplies and such like.
Norway’s Military Penal Code (1902), as amended in 1981, provides:
Anyone who contravenes or is accessory to the contravention of provisions relating to the protection of persons or property laid down in … the two additional protocols to [the 1949 Geneva] Conventions … is liable to imprisonment.
In 1987, in the Petane case, the Cape Provincial Division of South Africa’s Supreme Court dismissed the accused’s claim that the 1977 Additional Protocol I reflected customary international law. The Court stated:
The accused has been indicted before this Court on three counts of terrorism, that is to say, contraventions of s 54(1) of the Internal Security Act 74 of 1982. He has also been indicted on three counts of attempted murder.
The accused’s position is stated to be that this Court has no jurisdiction to try him.
… The point in its early formulation was this. By the terms of [the 1977 Additional] Protocol I to the  Geneva Conventions the accused was entitled to be treated as a prisoner-of-war. A prisoner-of-war is entitled to have notice of an impending prosecution for an alleged offence given to the so-called “protecting power” appointed to watch over prisoners-of-war. Since, if such a notice were necessary, the trial could not proceed without it, Mr Donen suggested that the necessity or otherwise for giving such a notice should be determined before evidence was led. …
On 12 August 1949 there were concluded at Geneva in Switzerland four treaties known as the Geneva Conventions. …
South Africa was among the nations which concluded the treaties. … Except for the common art 3, which binds parties to observe a limited number of fundamental humanitarian principles in armed conflicts not of an international character, they apply to wars between States.
After the Second World War many conflicts arose which could not be characterised as international. It was therefore considered desirable by some States to extend and augment the provisions of the Geneva Conventions, so as to afford protection to victims of and combatants in conflicts which fell outside the ambit of these Conventions. The result of these endeavours was Protocol I and Protocol II to the Geneva Conventions, both of which came into force on 7 December 1978.
Protocol II relates to the protection of victims of non-international armed conflicts. Since the State of affairs which exists in South Africa has by Protocol I been characterised as an international armed conflict, Protocol II does not concern me at all.
The extension of the scope of art 2 of the Geneva Conventions was, at the time of its adoption, controversial. …
The article has remained controversial. More debate has raged about its field of operation than about any other articles in Protocol I. …
South Africa is one of the countries which has not acceded to Protocol I. Nevertheless, I am asked to decide, as I indicated earlier, as a preliminary point, whether Protocol I has become part of customary international law. If so, it is argued that it would have been incorporated into South African law. If it has been so incorporated it would have to be proved by one or other of the parties that the turmoil which existed at the time when the accused is alleged to have committed his offences was such that it could properly be described as an “armed conflict” conducted by “peoples” against a “ra[c]ist regime” in the exercise of their “right of self-determination”. Once all this has been shown it would have to be demonstrated to the Court that the accused conducted himself in such a manner as to become entitled to the benefits conferred by Protocol I on combatants, for example that, broadly speaking, he had, while he was launching an attack, distinguished himself from civilians and had not attacked civilian targets. …
… I am prepared to accept that where a rule of customary international law is recognised as such by international law it will be so recognised by our law.
To my way of thinking, the trouble with the first Protocol giving rise to State practice is that its terms have not been capable of being observed by all that many States. At the end of 1977 when the treaty first lay open for ratification there were few States which were involved in colonial domination or the occupation of other States and there were only two, South Africa and Israel, which were considered to fall within the third category of ra[c]ist regimes. Accordingly, the situation sought to be regulated by the first Protocol was one faced by few countries; too few countries in my view, to permit any general usage in dealing with armed conflicts of the kind envisaged by the Protocol to develop.
Mr Donen contended that the provisions of multilateral treaties can become customary international law under certain circumstances. I accept that this is so. There seems in principle to be no reason why treaty rules cannot acquire wider application than among the parties to the treaty.
Brownlie Principles of International Law 3rd ed at 13 agrees that non-parties to a treaty may by their conduct accept the provisions of a multilateral convention as representing general international law. …
I incline to the view that non-ratification of a treaty is strong evidence of non-acceptance.
It is interesting to note that the first Protocol makes extensive provision for the protection of civilians in armed conflict. …
In this sense, Protocol I may be described as an enlightened humanitarian document. If the strife in South Africa should deteriorate into an armed conflict we may all one day find it a cause for regret that the ideologically provocative tone of s 1(4) has made it impossible for the Government to accept its terms.
To my mind it can hardly be said that Protocol I has been greeted with acclaim by the States of the world. Their lack of enthusiasm must be due to the bizarre mixture of political and humanitarian objects sought to be realised by the Protocol. …
According to the International Review of the Red Cross (January/February 1987) No 256, as at December 1986, 66 States were parties to Protocol I and 60 to Protocol II, which, it will be remembered, deals with internal non-international armed conflicts. With the exception of France, which acceded only to Protocol II, not one of the world’s major powers has acceded to or ratified either of the Protocols. This position should be compared to the 165 States which are parties to the Geneva Conventions.
This approach of the world community to Protocol I is, on principle, far too half-hearted to justify an inference that its principles have been so widely accepted as to qualify them as rules of customary international law. The reasons for this are, I imagine, not far to seek. For those States which are contending with “peoples[’]” struggles for self-determination, adoption of the Protocol may prove awkward. For liberation movements who rely on strategies of urban terror for achieving their aims the terms of the Protocol, with its emphasis on the protection of civilians, may prove disastrously restrictive. I therefore do not find it altogether surprising that Mr Donen was unable to refer me to any statement in the published literature that Protocol I has attained the status o[f] customary international [law].
I have not been persuaded by the arguments which I have heard on behalf of the accused that the assessment of Professor Dugard, writing in the Annual Survey of South African Law (1983) at 66, that “it is argued with growing conviction that under contemporary international law members of SWAPO [South-West Africa People’s Organisation] and the ANC [African National Congress] are members of liberation movements entitled to prisoner-of-war status, in terms of a new customary rule spawned by the 1977 Protocols”, is correct. On what I have heard in argument I disagree with his assessment that there is growing support for the view that the Protocols reflect a new rule of customary international law. No writer has been cited who supports this proposition. Here and there someone says that it may one day come about. I am not sure that the provisions relating to the field of application of Protocol I are capable of ever becoming a rule of customary international law, but I need not decide that point today.
For the reasons which I have given I have concluded that the provisions of Protocol I have not been accepted in customary international law. They accordingly form no part of South African law.
This conclusion has made it unnecessary for me to give a decision on the question of whether rules of customary international law which conflict with the statutory or common law of this country will be enforced by its courts.
In the result, the preliminary point is dismissed. The trial must proceed.
In 2010, in the Boeremag case, South Africa’s North Gauteng High Court stated:
In Petane, … Conradie J found that the provisions of [the 1977 Additional] Protocol I are not part of customary international law, and therefore are also not part of South African law.
Referring to the fact that in December 1986 only 66 of the 165 States party to the Geneva Conventions had ratified Protocol I, the Court [in Petane] stated:
This approach of the world community to Protocol I is, on principle, far too half-hearted to justify an inference that its principles have been so widely accepted as to qualify them as rules of customary international law. The reasons for this are, I imagine, not far to seek. For those States which are contending with “peoples[’]” struggles for self-determination, adoption of the Protocol may prove awkward. For liberation movements who rely on strategies of urban terror for achieving their aims the terms of the Protocol, with its emphasis on the protection of civilians, may prove disastrously restrictive. I therefore do not find it altogether surprising that Mr Donen was unable to refer me to any statement in the published literature that Protocol I has attained the status of customary international law.
Important changes with respect to certain aspects applicable at the time of Petane have taken place. The ANC [African National Congress] has become South Africa’s ruling party and in 1995 ratified Protocol I. The total number of States that have ratified it, is now … 162.
This last aspect forms the basis on which the First Respondent [the State] and the applicants agree that Protocol I forms part of customary international law as well as of South African law. As requested, this position is accepted for the purposes of the decision, without deciding on the matter.
Despite these changes, it remains debatable whether the provisions of Protocol I have become a part of South African law in this way.
The consensus of both parties to the conflict is required. See Petane … and Article 96 of Protocol I. …
Parliament’s failure to incorporate Protocol I into legislation in accordance with Article 231(4) of the Constitution in fact points to the contrary, and is indicative that the requirements of usus
and/or opinio juris
have not been met. See Petane
[footnotes in original omitted]
The Court also held:
If the [1977 Additional Protocol I] applies in South Africa as customary international law, the two requirements that form the basis of customary law must be met. It is arguable that the requirement of usus
has been met by the vast number of States that have acceded or ratified it. By ratifying Protocol I the Republic of South Africa has indicated its intention to apply the Protocol, thereby fulfilling the requirement of opinio juris
United States of America
In the Griffen case
in 1968, a US Army Board of Review confirmed the sentence of unpremeditated murder for having executed a Vietnamese prisoner, following a “manifestly illegal” order to do so. The accused declared that “he felt that the security of the platoon would have been violated if the prisoner were kept, since their operations had already been observed by another suspect”. The Board of Review cited paragraph 85 of the US Field Manual prohibiting the killing of prisoners of war. It added that the “killing of a docile prisoner taken during military operations is not justifiable homicide”.
Upon accession to the 1977 Additional Protocol I, Algeria stated that the term “feasible” must be interpreted as referring to “precautions and measures which are feasible in view of the circumstances and the information and means available at the time”.
Upon ratification of the 1977 Additional Protocol I, Belgium declared:
In view of the travaux préparatoires … “feasible precautions” [are] those that can be taken in the circumstances prevailing at the moment, which include military considerations as much as humanitarian ones.
At the CDDH, Canada stated that the word “feasible” when used in the 1977 Additional Protocol I, for example in Articles 57 and 58, “refers to what is practicable or practically possible, taking into account all circumstances existing at the relevant time, including those circumstances relevant to the success of military operations”.
Upon ratification of the 1977 Additional Protocol I, Canada stated: “The word ‘feasible’ means that which is practicable or practically possible, taking into account all circumstances ruling at the time, including humanitarian and military considerations.”
Upon ratification of the 1977 Additional Protocol I, France stated that it considered that the term “feasible” as used in the Protocol meant “that which can be realized or which is possible in practice, taking into account all circumstances ruling at the time, including humanitarian and military considerations”.
Upon ratification of the 1977 Additional Protocol I, Germany stated that it understood the word “feasible” to mean “that which is practicable or practically possible, taking into account all circumstances ruling at the time, including humanitarian and military considerations”.
Upon ratification of the 1977 Additional Protocol I, Ireland stated: “The word ‘feasible’ means that which is practicable or practically possible, taking into account all circumstances at the time, including humanitarian and military considerations.”
Upon ratification of the 1977 Additional Protocol I, Italy declared: “The word ‘feasible’ means that which is practicable or practically possible, taking into account all circumstances ruling at the time, including humanitarian and military considerations.”
At the CDDH, the Netherlands stated:
The word “feasible” when used in Protocol I, for example in Articles 50 and 51 [57 and 58], should in any particular case be interpreted as referring to that which was practicable or practically possible, taking into account all circumstances at the time.
Upon ratification of the 1977 Additional Protocol I, the Netherlands declared that “the word ‘feasible’ is to be understood as practicable or practically possible taking into account all circumstances ruling at the time, including humanitarian and military considerations”.
Upon ratification of the 1977 Additional Protocol I, Spain interpreted the term “feasible” as meaning that “the matter in question is feasible or possible in practice, taking into account all the circumstances prevailing at the time, including humanitarian and military aspects”.
United Kingdom of Great Britain and Northern Ireland
Upon signature of the 1977 Additional Protocol I, the United Kingdom stated: “The word ‘feasible’ means that which is practicable or practically possible, taking into account all circumstances at the time including those relevant to the success of military operations.”
United Kingdom of Great Britain and Northern Ireland
Upon ratification of the 1977 Additional Protocol I, the United Kingdom stated that it understood the term “feasible” as used in the Protocol to mean “that which is practicable or practically possible, taking into account all circumstances ruling at the time, including humanitarian and military considerations”.
United Kingdom of Great Britain and Northern Ireland
The Report on UK Practice cites a former director of the UK Army Legal Services who stated that UK soldiers were not required to risk their own lives in granting quarter. He added that it may not be practicable to accept surrender of one group of enemy soldiers while under fire from another enemy position. Capture was to take place when circumstances permitted.
United States of America
At the CDDH, the United States stated:
The word “feasible” when used in draft Protocol I, for example in Articles 50 and 51 [57 and 58], refers to that which is practicable or practically possible, taking into account all circumstances at the time, including those relevant to the success of military operations.
Diplomatic Conference on the Reaffirmation and Development of International Humanitarian Law Applicable in Armed Conflicts
The Report of Committee III of the CDDH stated:
Paragraph 3 [of Article 41 of the 1977 Additional Protocol I] dealing with the release of prisoners who could not be evacuated proved quite difficult. The phrase “unusual conditions of combat” was intended to reflect the fact that that circumstance would be abnormal. What, in fact, most representatives referred to was the situation of the long distance patrol which is not equipped to detain and evacuate prisoners. The requirement that all “feasible precautions” be taken to ensure the safety of released prisoners was intended to emphasize that the detaining power, even in those extraordinary circumstances, was expected to take all measures that were practicable in the light of the combat situation. In the case of the long distance patrol, it need not render itself ineffective by handing the bulk of its supplies over to the released prisoners, but it should do all that it reasonably can do, in view of all the circumstances, to ensure their safety.
Bothe, Partsch and Solf
In their commentary on the 1977 Additional Protocols, Bothe, Partsch and Solf mention Articles 19 and 20 of the 1949 Geneva Convention III (which require the prompt and humane evacuation of prisoners of war from the combat zone to places out of the danger area) and underline that “in certain types of operations, particularly airborne operations, commando raids, and long range reconnaissance patrols, compliance with these articles is clearly impractical, and there has been dispute as to what is required in such cases”.