While cluster bomb strikes endangered primarily populated areas, the aftereffects of cluster bombs caused more widespread harm to Afghanistan's civilians. Human Rights Watch found that BLU duds killed or injured scores of civilians. At least 127 casualties have been reported across eleven provinces. The majority of the victims were children. Unexploded cluster bombs also interfered with economic recovery, refugee repatriation, and military operations.
The precise dud, or initial failure, rate of cluster bombs used in Afghanistan is not known. Estimates range from 5 percent by the U.S. Air Force to up to 22 percent by deminers.143 In Kosovo, the U.N. Mine Action Coordination Center found that BLU-97 bomblets had a failure rate of about 7 percent.144 In order to calculate the precise dud rate, one must compare the number of unexploded bomblets with how many CBUs were dropped in a particular location. While deminers can eventually determine the number of unexploded bomblets, the United States has not provided the numbers of CBUs it used at each strike site in Afghanistan. The latter figure is difficult to determine on the ground in Afghanistan because people quickly gather the casings to sell as scrap metal in the bazaar.
Civilian Casualties from Duds
While antipersonnel landmines and other types of unexploded ordnance have caused far more casualties across Afghanistan,148 statistics show cluster bomblets can be more lethal. In Herat, between October 2001 and June 2002, unexploded cluster bomblets killed 44 percent of their victims while mines killed 21 percent. Other types of unexploded ordnance killed 41 percent of their victims.149
Shepherds, farmers, and children collecting firewood have been common victims in Afghanistan.150 Human Rights Watch learned of at least three shepherds killed by unexploded cluster bomblets in the Herat area. The strikes on Ishaq Suleiman left the village littered with BLUs. Demining consultant Moorhouse said there were BLUs "on houses, on roads, in gardens, in doorways, all over the place."151 Abdul Raziq, 43, and Ghouse-u-din, 37, brought their herds to a field west of the village four days after the bombing. As the men passed near the ruined shrine, a bomblet exploded and killed them both. "Before the explosion, people went there every day to graze, walk, use the fields. After that they avoided the area," said a cousin of the second victim.152 Deminers came to clear the site about one month later.153 In Shidai, nine miles (14.5 kilometers) east of Herat, a bomblet killed one shepherd and eight sheep, an OMAR team leader reported in late March.154 OMAR finished clearing that site in June.155
Given that Afghanistan is a largely agricultural country, farmers are also at high risk from unexploded BLUs. On the morning of December 21, for example, Arbrabrahim, 52, died while plowing a field on the north side of Jebrael with his oxen.156 The United States had bombed the village around October 28, presumably to attack the Taliban hiding there. Witnesses said Taliban troops had parked their trucks under trees at the edge of this field and then hid in local homes. Although the villagers had stayed in Jebrael during the Taliban occupation, no civilians died during the strike.157 The attack proved fatal after the fact, however. Other civilians have fallen victim to BLUs in Afghan farmland because crops are grown in small fields that are close to town and frequented by villagers. In Ishaq Suleiman, for example, 12-year-old Maroof died from an unexploded bomblet the day after the strike on the field southeast of town.158 Farmers face the greatest risks from these bomblets, which often sink into soft soil or hide in furrows, because they strike the submunitions forcefully and with metal tools.159
Cluster bombs have made gathering wood another dangerous occupation in Afghanistan. Afghans rely on wood for fuel, and they forage for it in the hills and rural areas outside their villages. In early December, 9-year-old Amin went to collect brush at the edge of the Jebrael field where Arbrabrahim had died. A cluster bomblet exploded and killed him.160 Children sent to gather wood also frequent military bases, which in Afghanistan generally have undefined and porous borders. The suburbs of Herat lost several children to bomblets dropped on these targets. Three children from Nawabad, for example, died while collecting wood at Firqa #17 in Herat.161
Children are especially vulnerable to cluster bomblets because of their curiosity. Arif, 14, and Sharif, 13, brothers from the Herat suburb of Bag Nazer Gah, were injured while playing during an excursion to the Firqa. Arif lost his leg to a bomblet.162 In nearby Qala Shater, Mohammed Eisah, 58, described how children played with BLUs before realizing their danger. "Two children were passing by. Other children misbehaved and threw a bomb[let] to their feet. Thank God, both survived," he said.163 Children have even interfered with demining efforts. "On the first day of [our clearance] work, children played among the bomblets," said a HALO Trust supervisor in the Shomali Plain. "We pushed them away. They shouldn't play there."164
New Year celebrations led to an increase in injuries to children and adults because people spent more time outside. On March 24, during the Human Rights Watch mission, five boys set off a cluster bomb while crossing a field in Takh-te-Sefar on their way to a picnic. Ramin, 15, died immediately. The other four boys were expected to survive but suffered injuries ranging from serious to minor. Soraj, 12, lost both legs. Ismaeel, 16, sustained a chest wound. Farhad, 18, injured his foot. Waheed, 5, received a chest wound and minor head injury.165 The cause of the explosion remains unclear. U.N. officials blamed a subsurface BLU, while the victims' relative, Ghulam Syed Siddiqi, 28, said one of the boys picked up the bomblet.166 Either way, the incident demonstrates the lingering dangers of cluster bombs to civilians.
Unexploded bomblets presented two other significant risks in Afghanistan although it is unclear if they caused any casualties. First, the collection of scrap metal from bombs put civilians in harm's way. "You never find the [CBU] casing. It goes to market and gets used for various products, like satellite dishes," said Bob Gannon, a demining consultant for Ronco.167 The casings themselves will not explode, but walking through fields to collect them or gathering canisters for scrap metal could set off unexploded bomblets. Human Rights Watch did not identify any people who were hurt in this way but found ample evidence of scrap gathering.168
Second, cluster bomb opponents expressed outrage during the war at the similarities in appearance between cluster bomblets and humanitarian daily rations.169 Although the latter are square and the former cylindrical, both are dropped from U.S. planes and both are yellow so that they are easy to spot on the ground. The United States responded to the outcry by changing the color of future food packages and warning civilians through announcements and flyers.170 Pentagon officials also said that the packages had not been dropped in the same areas as cluster bombs. "[D]espite the similarity in colors, it's very, very unlikely that a person would pick up a cluster bomb thinking it was a packet of food," Rear Adm. Craig Quigley, Deputy Assistant Secretary of Defense for Public Affairs, said.171 Human Rights Watch found no cases of people injured because they confused the two. Sean Moorhouse, the only deminer to note a correlation during the Human Rights Watch mission, said he had heard that some children thought that soda-can shaped BLUs were drinks to go with the food packages.172 Human Rights Watch did notice that yellow items proliferate in Afghanistan. Most families carry water in yellow plastic bottles. The omnipresent bottles and pieces are easily confused with BLUs, which demonstrates the danger of having unexploded bomblets lying around.
Although cluster bomb clearance is progressing well (see next chapter), casualties still occur. The ICRC reported twenty-one civilian casualties, including five deaths, since the Human Rights Watch mission in March 2002.173 The most recent reported casualty was a fifteen-year-old boy who suffered head and arm injuries on October 12 when a cluster bomblet exploded while he tended animals in the Oruzgan province.174 Two deminers from HALO Trust were also killed by cluster bomblets in the Kunduz province in 2002, one in Kunduz city in July, the other in Khoja Ghar in August.175
Cluster bomblets also hinder the return of refugees and internally displaced persons (IDPs). Afghanistan suffers from a long-standing refugee problem, which the most recent conflict aggravated.179 Between March and October 2002, the U.N. High Commissioner for Refugees (UNHCR) helped facilitate the repatriation of 1.6 million Afghans, and more than 400,000 people returned on their own.180 As of October, a year after the U.S. bombing began, however, more than two million refugees remained in the countries bordering Afghanistan.181 Bomblets and other unexploded ordnance can contribute to delays in organized returns. UNHCR's guidelines require the agency to look at the safety of roads and return areas before sending inhabitants back to their homes.182 The guidelines focus on the threats from landmines, but because of their similar effect, unexploded BLUs must be treated the same way.183 UNHCR is also required to discourage spontaneous repatriation in unsafe circumstances. "The need for return `in safety and dignity' means that UNHCR cannot promote the voluntary repatriation of refugees in patently dangerous situations with the risk of injury or death."184 By necessitating such precautions, cluster bombs can slow a country's economic recovery and its people's return to normalcy.
In Afghanistan, however, repatriation has happened quickly, which has increased concerns about safety. The rate of returns surprised most experts and caught agencies unprepared. As soon as the Taliban fled, Afghans started going back to their abandoned villages, some of which had been attacked with cluster bombs. In the Shomali Plain, as of March 2002, about ten of one hundred families had returned to Denar Khail, one of the communities most littered with unexploded BLUs,185 and twenty-five of 560 to Sabz Sang, where ten bomblets were observed laying in just one of its many small vineyards.186 HALO Trust was still doing active cluster clearance in both villages and ultimately destroyed 281 bomblets from the former and 208 from the latter.187 By late March, most villagers had returned to Mandisar, south of Kandahar, where BLUs lay among plants in a vineyard. About 115 people were using the land, and one man was injured when he picked up a bomblet with a spade.188
Unexploded bomblets also endanger transients unfamiliar with a region's hazards. Two people from the Mazlach IDP camp encountered cluster bomblets while passing through the field west of Ishaq Suleiman; the 61-year-old father died and his 8-year-old son was injured.189 Although the earlier deaths of two shepherds kept locals away, the victims had no reason to know of the incident. The deadly bomblets not only harm returning refugees but also contribute to a cycle of displacement, forcing those who find their villages too dangerous to join Afghanistan's large number of IDPs.
Conclusions and Recommendations
Human Rights Watch calls on the United States and other countries to discontinue the use of cluster bombs until they develop a submunition with a very low failure rate. The rate should certainly be less than 1 percent, and if technologically possible, considerably less than that. According to some experts, existing technology could achieve a failure rate of .1 to .3 percent.193
The U.S. government has adopted 1 percent or less as a reasonable goal for future production. On January 10, 2001, then-U.S. Secretary of Defense William Cohen issued a memorandum stating that it was the Defense Department's policy to "reduce overall UXO [unexploded ordnance] through a process of improvement in submunition system reliability-the desire is to field future submunitions with a 99% or higher functioning rate. . . . The Services shall design and procure all future submunition weapons in compliance with the above policy."194 The United States should be commended for recognizing the dangers of the excessive unexploded ordnance caused by cluster bomblets and other submunitions and for taking steps to rectify the problem. The new policy, however, permits continued use of existing submunitions that do not meet the new standard. Cohen said, "The services may retain `legacy' submunitions until employed or superceded by replacement systems . . . ." The United States stockpiles more than one billion submunitions with a failure rate of more than 1 percent.195 There is a fundamental inconsistency in acknowledging the dangers of these submunitions and the need to replace them while still permitting their use.
If the U.S. military cannot or does not modify the BLU-97 and other older submunitions to meet this standard, it should not employ them in any future conflicts. If it does decide to use them, it should restrict use of submunitions with a high failure rate to special circumstances where they are viewed as the only appropriate weapons for the mission and target. In any event, cluster bombs should not be used in or near populated areas or areas to which civilians are likely to return post-conflict.
While several types of technical approaches to reducing the dud rate are possible, Human Rights Watch is not in a position to evaluate or make recommendations in this regard. After nine months of research and consultations with munitions experts, the Vietnam Veterans of America Foundation recommended the incorporation of dual-event fuzes with a backup self-destruct mechanism.196 One way to fund such improvements would be to reallocate the cost of submunition parts. The shaped charge is the most expensive part of the $50 BLU-97. Human Rights Watch, however, found no evidence of cluster bombs being used mainly as an anti-armor weapon in Afghanistan.197 It has been suggested that if the BLU were recognized as a primarily antipersonnel weapon and the shaped charge were removed, more resources could be directed to developing a fuze that would lower the dud rate to below 1 percent.198
The military is also capable of designing a unitary bomb that has the same antipersonnel effect as a cluster bomb. During the Afghanistan air war, the United States experimented with such an idea. It set Joint Direct Attack Munitions (JDAMs), a type of 2000-pound satellite-guided bomb, to burst in the air so that they would release fragments akin to the BLU's fragmentation core.199 A unitary antipersonnel bomb would raise the same targeting issues as a cluster bomb, but it would lessen concerns about aftereffects.
The cluster bomb dud problem can be attacked through targeting as well as technology. Users should avoid dropping cluster bombs from high altitudes or in certain environments, such as soft ground, because past experience shows that the dud rate is likely to increase in these situations. The impact of the dud problem can also be lessened greatly if states prohibit the use of cluster bombs in or near populated areas or in areas to which civilians may return en masse.
In sum, Human Rights Watch recommends that:
· Countries suspend the use of cluster bombs until the dud rate is reduced dramatically. By order of the Secretary of Defense, future U.S. submunitions must have a failure rate of at least less than 1 percent.
While solutions discussed above hold promise, one must remember that cluster bomb strikes would still raise targeting issues even if their dud rate were eliminated.
143 HALO Trust said it estimated dud rates ranging from 15 to 22 percent. Farnaz Fassihi, "Death Lurks Underfoot," Star-Ledger (Newark), December 23, 2001.
144 ICBL, Landmine Monitor Report 2001 (New York: Human Rights Watch, 2001), p. 952.
145 ICRC, Mine Victims Report, October 6, 2001 to date [hereinafter ICRC, November Cluster Bomb Casualty List]. Human Rights Watch obtained this document, which lists only cluster bomb casualties, in November 2002. A list from June 2002 reported eight-seven casualties, including thirteen deaths, from unexploded cluster bomblets. ICRC, "Number of Cluster Munition Victims Recorded Since October 2001 to June 2002," June 22, 2002. An earlier list in March 2002 reported fifty-nine casualties, including nine deaths. ICRC, "All Afghanistan Cluster Ammunition Casualties," October 2001-March 2002.
146 The provincial breakdown was as follows: Nangarhar, forty-seven victims; Herat, thirty-three; Kandahar, thirteen; Kabul, eleven; Bamiyan, eight; Baghlan, Hilmand, Parwan, Kunduz, three each; Uruzgan, two; Zabul, one. ICRC, November Cluster Bomb Casualty List.
147 Human Rights Watch interview with Mohammed Kazim Malwan Ahmadzai, Deputy Program Manager of the Mine Data Collection Program, ICRC, Kabul, Afghanistan, March 12, 2002. Moorhouse concurred, saying, "A lot of people who died are buried [immediately] so they are not reported. [Deaths] are only reported if someone is killed and someone is injured [because the injured will report the death to the hospital]. Deaths are underreported." Human Rights Watch interview with Sean Moorhouse.
148 According to the ICRC, in calendar year 2001, landmines caused 472 casualties, unexploded ordnance 476 casualties, antivehicle mines thirty-five casualties, booby-traps fourteen casualties, fuzes fifty casualties, and cluster munitions sixty-three casualties. Of the sixty-three cluster munition casualties, forty-eight occurred between October and December 2001. ICBL, Landmine Monitor Report 2002 (New York: Human Rights Watch, 2002), p. 603.
149 Cluster bomblets killed twenty civilians and injured twenty-five. Other types of unexploded ordnance killed five and injured seven. Mines killed three and injured eleven. RMAC Herat, Cluster Victims 2002; RMAC Herat, Mine/UXO Victims 2002.
150 The ICRC list corroborates the trends Human Rights Watch identified during its mission to Afghanistan. Of the victims it reported, 20 percent were tending animals, 16 percent were farming, and 10 percent were gathering wood when injured. The list breaks down the victims' activities at the time of incident as follows: tending animals, 25 victims; farming, 20; traveling on foot, 19; playing/recreation, 15; collecting wood, 13; incidental passing, 13; tampering with item, 9; traveling in vehicle, 2; military activities, 2; other, 7; unknown, 2. ICRC, November Cluster Bomb Casualty List.
151 Human Rights Watch interview with Sean Moorhouse.
152 Human Rights Watch interview with Abdul Basir and Shames-u-din. Abdul Basir, 37, was the brother-in-law of Abdul Raziq. Shames-u-din, 31, was the cousin of Ghouse-u-din.
154 Briefing by OMAR team leader, Shidai, Afghanistan, March 28, 2002.
155 OMAR, Adopt-a-Team Monthly Progress Report, June 2002.
156 Human Rights Watch interview with Abdul Naim, Jebrael, Afghanistan, March 29, 2002. Abdul Naim, a 25-year-old farmer, was the victim's son. Other witnesses confirmed his story. See also Human Rights Watch interview with Ali, 12-year-old carpet weaver, Jebrael, Afghanistan, March 29, 2002; Human Rights Watch interview with Abdul Khaliq, Jebrael, Afghanistan, March 29, 2002. Abdul Khaliq, 30, was the victim's nephew.
157 Human Rights Watch interview with Ali, 38-year-old soldier, Jebrael, Afghanistan, March 29, 2002. "At the time the village was bombed, the Taliban were there. They brought their vehicles under the trees and in these areas [pointing to a nearby street] and therefore it was bombed. . . . They hid under where we make bricks. They also hid in houses. People protested. They left vehicles by houses and left the area. Later the Taliban came and took their vehicles. We had great fear of the Taliban and couldn't get closer to them," Ali said.
158 Human Rights Watch interview with Khalil Ahmad.
159 Landmine Action, Explosive Remnants of War, p. 29.
160 Human Rights Watch interview with Ali, 12-year-old carpet weaver.
161 Human Rights Watch interview with Maidin, Nawabad, Afghanistan, March, 29, 2002. Maidin, 18, said he knew of two additional local civilians injured by bomblets on the Firqa.
162 Human Rights Watch interview with Sharif, Bag Nazer Gah, Afghanistan, March 29, 2002.
163 Human Rights Watch interview with Mohammed Eisah.
164 Human Rights Watch interview with Habeeb, Explosive Ordnance Disposal (EOD) Supervisor, HALO Trust, Shomali Plain, Afghanistan, March 14, 2002.
165 Human Rights Watch interview with Ghulam Syed Siddiqi, Gazer Gah, Afghanistan, March 29, 2002. Ghulam Syed Siddiqi, 28, was the cousin of the boy who died and either uncle or cousin to the other boys.
167 Human Rights Watch interview with Bob Gannon. See also Human Rights Watch interview with HALO Trust team leader, Sabz Sang, Afghanistan, March 14, 2002.
168 For evidence of scrap gathering despite the risks in Ishaq Suleiman, see text following footnote 246. OMAR's list of cluster bomb casualties includes four injuries at Firqa #17 and four at the Qol-e Urdu that happened while people were "collecting scrap and tending animal[s]." OMAR Sub Office Herat, "Information About Victims" [hereinafter OMAR Cluster Bomb Casualty List].
169 See, e.g., Human Rights Watch, "Cluster Bombs in Afghanistan," p. 3; Steven Mufson, "Pentagon Changing Color of Airdropped Meals," Washington Post, November 2, 2001; "Human Rights Groups Are Protesting Use of Cluster Bombs," CNN: Live This Morning transcript, November 1, 2001.
170 "DoD News Briefing-Secretary Rumsfeld and Gen. Myers," U.S. DoD News Transcript, November 1, 2001. Air Force communications planes issued radio broadcasts in the local languages, Pashto and Dari, warning of the dangers. The broadcasts said, "We do not wish to see an innocent civilian mistake the bombs for food bags and take it away, believing it might contain food. . . . Do not confuse the cylinder-shaped bomb with the rectangular food bag." Another broadcast warned, "Attention, people of Afghanistan! . . . Let the bundles [of rations] land and settle before you approach them. If you follow these instructions you will not get hurt." Richard Sisk, "U.S. Warns of Food, Bomb Mixups," New York Daily News, October 31, 2001.
171 "Rear Admiral Quigley Briefing at the Foreign Press Center," U.S. DoD News Transcript, November 14, 2001.
172 Human Rights Watch interview with Sean Moorhouse.
173 ICRC, November Cluster Bomb Casualty List.
175 Human Rights Watch interview with Abdul Wasi, HALO Trust, Kabul, Afghanistan, November 2002. See also Mine Action Program for Afghanistan (MAPA), Chart of Cluster Bomb Clearance and Casualties. Human Rights Watch obtained this document from MAPA in November 2002.
176 Human Rights Watch interview with Karlwan.
177 Human Rights Watch interview with Gerhard Zenk, Kunduz.
178 Human Rights Watch interview with Gerhard Zenk, Pul-i Khomri.
179 The September 11 attacks and fear of U.S. bombing led a new wave of Afghans to flee the country. Within ten days of the terrorist attacks on New York and Washington, 20,000 fled into Pakistan and Iran. Human Rights Watch, "Safe Refuge Must Be Provided for Afghan Refugees," Press Release, September 21, 2001.
180 Ruth Gidley, "Agencies Caught Off Guard as Afghans Flock Home," Reuters, October 21, 2002 (quoting UNHCR figures). While these numbers come from UNHCR, they may be exaggerated. A London Sunday Telegraph article suggested that many of those who returned did so several times in order to collect relief packages. Christina Lamb, "Afghan Refugees Run a Scam on U.N. Relief," London Sunday Telegraph, November 11, 2002.
181 UNHCR, "Afghanistan at a Glance," at http://www.unhcr.ch/cgi-bin/texis/vtx/afghan?page=facts (last visited October 11, 2002).
182 Handbook: Voluntary Repatriation: International Protection (Geneva: UNHCR, 1996), sec. 6.5 ("The presence of landmines and other unexploded ordnance on main routes of return and in returnee settlement areas poses tremendous danger for repatriating refugees and is therefore a major protection concern to UNHCR."). UNHCR recognizes that not every mine "poses an immediate threat to voluntary repatriation," so it focuses its study on "the routes of organized as well as, to the extent possible, spontaneous return, and the sites where refugees will settle upon their return." Ibid.
183 UNHCR has recognized the dangers of unexploded ordnance, such as cluster bomblets. "Afghanistan's challenges remain daunting. Insecurity, along with mines and unexploded ordnance, still affect some areas . . . ." UNHCR, "Afghanistan at a Glance."
184 Handbook: Voluntary Repatriation: International Protection, sec. 6.5 (emphasis in original). See also UNHCR, Handbook for Emergencies, 2nd. ed., p. 282.
185 Human Rights Watch interview with HALO Trust team leader, Denar Khail, Afghanistan, March 14, 2002.
186 Human Rights Watch interview with HALO Trust team leader, Sabz Sang.
187 RMAC Kabul, "BLU Strick (sic) Sites Report as at 31 Aug 02" [hereinafter RMAC Kabul, "BLU Strike Sites"]. This document represents the information HALO Trust provided to the RMAC.
188 Human Rights Watch interview with Haji Faizul Haq, Team Leader #4, Demining Agency for Afghanistan (DAFA), Mandisar, Afghanistan, March 24, 2002.
189 Human Rights Watch interview with Abdul Basir and Shames-u-din. See also OMAR Cluster Bomb Casualty List.
190 "CBS Evening News," CBS News Transcript, March 18, 2002. According to another report, "[t]he soldiers described the area as being littered with U.S. cluster bombs and unexploded ordnance, adding to the dangers faced by troops as they searched the peak." Stephen Coates, "Al-Qaeda Cave Stronghold Was Like a Castle, Say U.S. Troops," Agence France-Presse, March 16, 2001.
191 "CBS Evening News."
192 See footnotes 47-49 and accompanying text.
193 See Vietnam Veterans of America Foundation, Proposed Protocol To Address Explosive Remnants of War, September 25, 2001, slide 10.
194 Secretary of Defense William Cohen, Memorandum for the Secretaries of the Military Departments, Subject: DoD Policy on Submunition Reliability (U), January 10, 2001. The memo defines future submunitions as those reaching a production decision in fiscal year 2005 and beyond. The memo also notes that "functioning rates may be lower under operational conditions due to environmental factors such as terrain and weather."
195 Human Rights Watch, A Global Overview of Explosive Submunitions, p. 2.
196 Vietnam Veterans of America Foundation, Proposed Protocol, slide 10.
197 The statement of an Air Force spokesman corroborates Human Rights Watch's finding. The CBU-103 is "useful primarily against softer targets--troops, softer vehicles, anything that's not armored basically. . . ." Rich Tuttle, "Wind Corrected Munitions Dispenser in First Combat Use, Program Office Director Says," Aerospace Daily, December 5, 2001.
198 This idea was proposed by William M. Arkin, senior military advisor to Human Rights Watch and adjunct professor at the U.S. Air Force's School of Advanced Airpower Studies.
199 JDAMs can be set to "air-burst, impact or penetrating modes." David Fulghum and Robert Wall, "Heavy Bomber Attacks Dominate Afghan War," Aviation Week & Space Technology 155 (December 3, 2001). See also Bryan Bender, Kim Burger, and Andrew Koch, "Afghanistan: First Lessons," Jane's Defence Weekly (December 19, 2001) (discussing the criticism of cluster bombs, this article notes that "in recent years the DoD has begun developing unitary warheads for many of its strike weapons, at least partially due to pressure over submunition use.").