This сase presents a question of first impression in this circuit: as a 'matter of constitutional law, can a school district, acting in good faith in an effort to deal with a serious drug and alcohol problem, subject students, their lockers, and their automobiles to the exploratory sniffing of dogs trained to detect certain contraband? We must consider the special circumstances peculiar to the public school environment, the duty of school officials to protect the minors in their care, the growing problem of drug and alcohol abuse in the schools, the students’ interest in the integrity of their persons and effects, and the importance of demonstrating to the young that constitutional guarantees are not only lofty theories but do in practice control our government. Bearing in mind all these considerations, we hold that the dogs’ sniffing of the children was unconstitutional. We conclude, however, that the dogs’ sniffing of cars and lockers does not rise to the same level of intrusiveness, and we hold that, in the school environment, such sniffing operations are permissible.
I.
The named plaintiffs, Robby Horton, Heather Horton, and Sandra Sanchez, brought this action by their next friend, Robert Horton, seeking to represent all students enrolled in the Goose Creek Consolidated Independent School District (GCISD) in a challenge under 42 U.S.C. § 1983 of the defendant school district’s canine drug detection program.
The defendant, GCISD, adopted the challenged program in response to a growing drug and alcohol abuse problem in the schools. It contracted with a security services firm, Securities Associates International, Inc. (SAI), that provides dogs trained to alert their handlers to the presence of any one of approximately sixty different substances, including alcohol and drugs, both over-the-counter and controlled. The defendant conducted assemblies in the elementary schools to acquaint the children with the dogs and informed students in the junior and senior high schools of the program. On a random and unannounced basis, the dogs are taken to the various schools in the district, where they sniff students’ lockers and automobiles. They also go into the classrooms, on leashes, to sniff the students themselves. During their “playtime” at the schools, the dogs are sometimes taken off their leashes. When a dog alerts the handler to the odor of an illicit substance on a student’s person, after the sweep of the class is completed and the dog and handler have departed, a school official discreetly asks the student to leave the class and go to the administrator’s office, where he is subjected to a search of pockets, purse, and outer garments.
The named plaintiffs were all subjected to the sniffing of the canine drug detectors. Two of them, Robby Horton and Sandra Sanchez, triggered alerts. School officials questioned Sandra, took her purse, and searched it without her consent. They found a small bottle of perfume, which they returned to her. Robby was asked to empty his pockets, which he did. When nothing incriminating was found, the school officials searched his socks and lower pants legs but again found no contraband.
The plaintiffs brought this action, alleging a violation of the fourth amendment prohibition of unreasonable searches and seizures and a violation of the fourteenth amendment prohibition of deprivations of liberty and property without due process. On a motion for class certification and cross-motions for summary judgment, the district court denied certification and held that the sniffing, although it is a search, is not unrеasonable. Further, it held that reasonable cause is the standard for searches of students and their property by school officials acting in loco parentis, and the alert of the dogs provides reasonable cause for searches of lockers and cars as well as for searches of the pockets, purses, and outer garments of students. Finally, the district court held that the program does not violate the due process clause, because it subjects the students to minimal intrusion, humiliation, and fear. The plaintiffs appeal both on the merits and on the question of class certification.
II.
Although the problem the merits present in this case is new to the Fifth Circuit, a district court in this circuit and appellate courts for the Seventh and Tenth Circuits have decided similar cases. In the most recent case, Zamora v. Pomeroy, 10 Cir. 1981,
The Seventh Circuit reached the same result on facts similar to those presented by the GCISD program. In Doe v. Renfrow, N.D.Ind.1979,
A district court in our own circuit, on the other hand, reached the opposite result, explicitly rejecting Doe v. Renfrow. Jones v. Latexo Independent School District, E.D. Tex.1980,
The problem presented in this case is the convergence of two troubling questions. First, is the sniff of a drug-detecting dog a “search” within the purview of the fourth amendment? Second, to what extent does the fourth amendment protect students against searches by school administrators seeking to maintain a safe environment conducive to education? On each question, we find an abundance of precedent but scant guidance.
A. The Canine Sniff as a Search
Frequent use of drug-detecting dogs by law enforcement officials has led to a great number of cases challenging the admissibility of the fruits of a canine sniff.
Both the general characteristics of a canine sniffing investigation and the specific features of the object of the investigation— car, locker, or person — are relevant in determining the constitutionality of the search.
We find this reasoning unpersuasive. The sniffing of a dog is unquestionably “different” from the sniffing of a human being; otherwise a rational law enforcement agency would not invest resources in training the animals but would simply send the handlers to sniff. The trained dogs— occasionally described as “giant olfactory nerve[s]”
The application of these principles to the sniff of a drug-detecting dog suggests that there was a search in this case. The trained dog can dеtect odors that no human investigator could possibly perceive. And the use of such dogs is not widespread in society; we cannot say that the revelation of these undetectable odors is a concomitant of life in society.
In addition to these considerations, though, we must evaluate also the nature of the object of the investigation.
The use of the dogs to sniff the students presents an even clearer instance of a search within the fourth amendment. The students’ persons certainly are not the subject of lowered expectations of privacy. On the contrary, society recognizes the interest in the integrity of one’s person, and the fourth amendment applies with its fullest vigor against any indecent or indelicate intrusion on the human body.
Our decision that the procedures employed by GCISD constitute searches does not, however, compel the conclusion that they were constitutionally impermissible. The fourth amendment does not prohibit all searches; it only restricts the government to “reasonable” searches. The reasonableness of these procedures turns partly on the school environment, to be discussed in Part IIB. But the reasonableness is also governed in part by general fourth amendment principles.
The body of case law on canine sniffing is best described as the implicit recognition by the courts of a new category under the fourth amendment — the “limited search”, similar to that recognized by the Supreme Court in Terry v. Ohio, 1968,
Although we recognize significant drawbacks to the adoption of a “sliding scale” approach to fourth amendment analysis,
Similarly, when the object of the search is a person, the balance permits the search only if there is a reasonable suspicion.
The preceding discussion, however, does not resolve the case currently before us. We have established that dragnet searches of persons and property by sniffing dogs are ordinarily prohibited. But the fourth amendment does not always require the same results in the schools as it does in ordinary circumstances. As a result, we must consider the extent to which the public school setting affects fourth amendment analysis.
B. The Fourth Amendment in the Public Schools
The courts have encountered substantial difficulty in accommodating the fourth amendment to the special situation presented by the public schools, where school officials have both a right and a duty to provide a safe environment conducive to education. At one time, it was not uncommon for a court to view the school official who searched a student as acting under authority derived from the parent and therefore as a private party not subject to the constraints of the fourth amendment. See, e.g., Mercer v. State, Tex.Civ.App.—Austin 1970,
But the decision that school officials are governed by the fourth amendment does not dictate a holding that their activity in this case was unconstitutional. The basic concern of the fourth amendment is reasonableness,
At least one case has held that the reasonable cause standard applicable in thе schools requires individualized suspicion. Bellnier v. Lund, N.D.N.Y.1977,
None of the administrative search cases, though, involved searches of persons, which entail a far greater invasion of privacy. In Camara, the Court explicitly noted as support for its decision that the search was not “personal in nature [and therefore] involve[d] a relatively limited invasion of the ... citizen’s privacy”.
One hurdle remains to the validation of the dragnet sniffing of lockers and cars— the failure to obtain a warrant. The usual fourth amendment rule is that, even if supported by probable cause, a warrantless search is unreasonable, subject to a few well-delineated exceptions. See, e.g., Katz v. United States, 1967,
In the schools, however, as discussed above, blanket rules are of little use, for the proper inquiry is the overall reasonableness of the search.
The plaintiffs urge that, even if the initial sniffing of the cars and lockers by the dogs is permissible, the dogs’ reactions do not give the defendant a sufficiently strong basis for suspicion to justify a further search. The district court stated that the “generalized perception of a problem of drug and alcohol abuse” along with the positive reaction of the dog gave the school sufficient cause to believe that the student occupant or driver had violated school policy to justify opening the locker or car and searching it. The court did not, however, make any finding on the reliability of the dogs, and there was no evidence in the record to support such a finding. In fact, although the representative of the SAI asserted that the dogs were quite reliable, he admitted that there were no comprehensive records kept of those incidents when the dogs reacted positively in the absence of contraband. On this record, then, we cannot say whether the reaction of the dogs provided adequate cause for more intrusive searches, and summary judgment is inappropriate. Fed.R.Civ.P. 56(c). We remand to the district court for development of the record on that point. The standard enunciated by the district court, however, was proper: GCISD need not show that the dogs are infallible or even that they are rеliable enough to give the defendant probable cause; instead, the dogs must be reasonably reliable.
III.
The plaintiffs also argue that the use of the dogs violates their rights under the fourteenth amendment, by depriving them of a liberty interest without due process. Because of our disposition of the fourth amendment issues arising out of the sniffing of the students, see Part II, we need not decide whether that practice entails a due process violation. The question remains whether the presence of a dog on campus and the practice of occasionally allowing him to play on campus, unrestrained by a leash but supervised by the handler, constitute a violation of the due process clause.
The dogs trained and provided by SAI are large animals — usually German shepherds and Doberman pinschers, and occasionally labradors — breeds selected because the animals are often sold to police forces, who wish to maintain an image of strength and ferocity. The individual animals, however, are selected on the basis of their docility, and SAI has never received a complaint about the dogs’ injuring anyone in any way
We recognize that large dogs, particularly those breeds that are sometimes used as attack dogs, often engender an irrational fear, and we do question the wisdom of permitting them to roam parts of the campus unleashed. But, as long as the dogs are carefully selected for their nonaggressive character, and the handlers supervise them during their playtime, we do not think that the minimal “harassment” arising from their mere presence on campus rises to the level of a constitutional violation.
IV.
The plaintiffs sought to maintain a class action under rule 23, Fed.R.Civ.P., requesting the district court to certify a class of all students currently enrolled in GCISD schools. The defendant opposed class certification, and the district court refused to certify the class.
The decision to grant or to deny certification is, as the defendant contends, initially committed to the sound discretion of the district judge, and the decision will not be overturned except for abuse of discretion. See, e.g., Doninger v. Pacific Northwest Bell, Inc., 9 Cir. 1977,
(1) The class must be so numerous that joinder of all members is impracticable;
(2) There must be questions of law or fact common to the class;
(3) The claims or defenses of the representative parties must be typical of the claims or defenses of the class; and
(4) The representative parties must fairly and adequately protect the interests of the class.
Fed.R.Civ.P. 23(a). In addition, the action must fit within one of the categories of actions described in rule 23(b).
The district court identified correctly the cause for concern here as the adequacy of the named plaintiff, for it is clear that the case satisfies the other prerequisites to a class action.
The adequacy requirement mandates an inquiry into the zeal and competence of the representative’s counsel and into the willingness and ability of the representative to take an active role in and control the litigation and to protect the interests of absentees, see, e.g., Jaurigui v. Arizona Board of Regents, D.Ariz.1979,
But the possibility of antagonism
We perceive unresolved tension between the cases permitting certification in these circumstances and the leading case on adequacy of representation for purposes of binding class members, Hansberry v. Lee, 1940,
It is one thing to say that some members of a class may represent other members in a litigation where the sole and common interest of the class in the litigation is either to assert a common right or challenge an asserted obligation. It is quite another to hold that all those who are free alternatively either to assert rights or to challenge them are of a single class so that any group merely because it is of the class so constituted, may be deemed adequately to represent any others of the class in litigating their interests in either alternative. Such a selection of representatives ... does not afford that protection to absent parties which due process requires.
Hansberry, however, cannot be read to forbid class actions
... [i]n any conceivable case, some of the members of the class will wish to assert their rights while others will not wish to do so. Thus the familiar case of thе stockholders’ derivative suit is almost invariably brought by minority stockholders to challenge action that a majority of the stockholders approve. Yet it is routinely regarded as an appropriate class suit. Another familiar class suit is that in which one or more taxpayers of a community, suing on behalf of all, challenge the validity of a proposed public expenditure. It is difficult to believe that there has ever been such a case in which a good many of the taxpayers would not have preferred that their rights not be enforced, because of their interest in having the expenditure made. Yet no one has ever doubted the propriety of bringing such a suit as a class action.
In this case, a variety of techniques was available to the district judge. For instance, he could have ordered notice of the action and of the relief requested by the plaintiff to the other students and parents to be posted or distributed to the students at the schools, an effective and relatively inexpensive way to apprise other members of the litigation and to invite intervention to challenge the representation or to oppose the named plaintiffs.
We think it unnecessary to undertake these procedures in this particular case, although their use might increase the protection of the absentees, for we think that the parties in this case protected the interests of all absentees, as required by Hansberry,
V.
We conclude thаt the use of dogs in dragnet sniff-searches of the students of GCISD is unconstitutional, and we direct the district court to grant relief by appropriate declaration and injunction. Although the use of the dogs in similar dragnet sniffing of lockers and cars is permissible, we must remand to the district court for the case to proceed to trial on the reliability of the dogs’ reactions as the basis for further searches. We also direct certification of a class on the issue of the constitutionality of the practices.
REVERSED AND REMANDED.
Notes
. The parties agree that no such intrusive searches as strip searches or body cavity searches occur.
. We use “contraband” to refer to all substances that the school forbids students to possess, even if possession violates no law.
. The following list is not exhaustive. United States v. Johnson, 2 Cir. 1981,
. See, e.g., United States v. Goldstein, 5 Cir. 1981,
. State v. Elkins, 1976,
. United States v. Venema, 10 Cir. 1977,
. United States v. Solis, 9 Cir. 1976,
. State v. Martinez, 1976,
. All the cases cited in note 3, with the exception of United States v. Burns, 10 Cir. 1980,
. Some cases permit dragnet canine sniffing for reasons largely unrelated to the use of dogs. For instance, in People v. St. George Matthews, 1980,
. In particular, all of the federal cases cited in note 2 show some degree of individualized suspicion, with the exception of the customs search in Race, see note 10, and possibly United States v. Johnson, 2 Cir. 1981,
Also, neither of the cases binding upon us involved dragnet searches. United States v. Viera, 5 Cir. 1981,
. The commentators seem unanimous in viewing canine sniffing as a search. See 1 W. LaFave, Search and Seizure § 2.2(f) at 282 n.166, 283, Supp. 51 n.185 (1978 & Supp.1982) (describing Wolohan as an “outrageous” result); Peebles, The Uninvited Canine Nose and the Right to Privacy: Some Thoughts on Katz and Dogs, 11 Ga.L.Rev. 75 (1976); Comment, Search and Seizure in the Public Schools: Are Our Children’s Rights Going to the Dogs? 24 St. Louis U.L.J. 119 (1979); Note, Police Use of Sense-Enhancing Devices and the Limits of the Fourth Amendment 1977, U.M.L.F. 1167, 1197— 1201; Note, Constitutional Limitations on the Use of Canines to Detect Evidence of Crime, 44 Fordham L.Rev. 973 (1976). One article is a bit more equivocal when property is concerned but concludes that sniffing of persons is a search. Gardner, Sniffing for Drugs in the Classroom— Perspectives on Fourth Amendment Scope, 74 Nw.U.L.Rev. 803 (1980).
. We reject the opinion expressed in Doe v. Renfrow that the knowledge of a drug problem in the schools is equivalent to the grounds for suspicion in the earlier cases — in effect, that the search in Doe and the search here are not dragnet searches.
. We reserve consideration of the special factors inherent in the school environment until Part IIB, below.
. See, e.g., United States v. Goldstein, 5 Cir. 1981,
. See, e.g., United States v. Bronstein, 2 Cir. 1975,
. Jones v. Latexo Independent School District, E.D.Tex.1980,
. United States v. Michael, 5 Cir. 1981,
. See United States v. Michael, 5 Cir. 1981,
Similarly, and closer to home, we do not think that it would be impermissible to use a dog to track a suspect. Although the human olfactory sense might not be sharp enough to pick up the scent, the investigators are not discovering data otherwise imperceptible to all human senses; human investigators could have traced the suspect’s movements by following him, using unaided vision.
. To this point, our discussion has focused on property, because the arguments have been raised in cases dealing with sniffing of property. See note 9. We see no reason why the open view concept, extended to “open smell”, should not apply to persons. Consequently, the argument that a dog’s sniff, reveals only information subject to “open smell” should be applicable (or inapplicable) to persons just as it is to property.
. The school district has not argued that it owns the lockers and permits the students to use them subject to a retained right of inspection. Although ownership alone does not determine the scope of fourth amendment protection, and announcements that privacy will not be respected do not, per se, preclude reasonable expectations of privacy, these factors can be relevant. Thus, in Zamora v. Pomeroy, 10 Cir. 1981,
It is the only place where [the student] may be able to store what he seeks to preserve as private — letters from a girl friend, applications for a job, poetry he is writing, books that may be ridiculed because they are too simple or too advanced, or dancing shoes he may be embarrassed to own.
Buss, The Fourth Amendment and Searches of Students in Public Schools, 59 Iowa L.Rev. 739, 772-73 (1974). That factor, along with the common practice of assigning each student exclusive use of a locker and treating it as his dominion and responsibility, we think, gives the student a reasonable expectation of some privacy in the contents of his locker.
. Cf. United States v. Viera, 5 Cir. 1981,
. In holding that canine sniffing is not a search, courts frequently note that the search is a minimal intrusion because the use of the dogs discriminates among types of information, and any error is in favor of the suspect. The handler discovers only whether the object contains contraband, unlike one who examines the contents of a container visually and views all innocent possessions as well as any contraband. That consideration ought to relate not to the categorization of the procedure as a “search” but to the reasonableness of the search. See Part IIB, below. At any rate, on this record, that consideration simply is not present. The defendant has introduced no evidence of the reliability of the dogs, while the plaintiff has shown two instances in which the error was in favor of further searches rather than in favor of the suspect. In addition, it is conceded that the dogs in this case alert to some sixty substances including anything that, like the bottle of perfume in Sandra Sanchez’s purse, contains alcohol. The handlers, therefore, obtain information about innocent рossessions as well as about violations of school policy.
. The record contains evidence that the dogs put their noses “up against” the objects or persons they are investigating. Newman Depo. at 43. We therefore reject the school district’s suggestion that the record shows no physical contact between the dogs and the students. In any event, physical trespass is no longer the key to distinguishing searches from non-searches, Katz v. United States, 1967,
. This result accords fully with all the precedent on canine sniffing except Doe v. Renfrow. Every other case concerned unattended property, see note 9, and courts frequently gave as one reason in favor of categorizing sniffing as a non-search that “[t]he invasion of privacy is not aimed at the person but rather his luggage. Thus, the scope of the intrusion is minimal, restrictive and tolerable by society.” Mata v. State, Fla.App.1980,
. See 1 W. LaFave, Search and Seizure § 2.2(f) at 286 (1978).
. Several commentators have explained the cases in this way. See 1 W. LaFave, Search and Seizure § 2.2(f) at 288 (1978); Peebles, The Uninvited Canine Nose and the Right to Privacy: Some Thoughts on Katz and Dogs, 11 Ga.L. Rev. 75 (1976); Note, Constitutional Limitations on the Use of Canines to Detect Evidence of Crime, 44 Fordham L.Rev. 973 (1976); Note, Police Use of Sense-Enhancing Devices and the Limits of the Fourth Amendment, 1977 U.Ill. L.F. 1167, 1168-69, 1201. A few cases have explicitly adopted the viewpoint that a canine sniff is a search but a reasonable one. State v. Elkins, 1976,
. See generally Amsterdam, Perspectives on the Fourth Amendment, 58 U.Minn.L.Rev. 349, 388-95 (1974); Peebles, The Uninvited Nose and the Right to Privacy: Some Thoughts on Katz and Dogs, 11 Ga.L.Rev. 75, 103-04 (1976).
. In most cases, the record has established that the dogs are overwhelmingly correct and that any errors favor the suspect. In this case, there is no evidence of reliability, and, since the dogs alert handlers to some sixty substances, including innocent substances, the information obtained by the handlers is not as limited. See note 23. Nonetheless, the handlers obtain much less information than is revealed by a visual examination.
. It is in this sense that the use of dogs is a commendable, intelligent, and responsible police procedure. See United States v. Fulero, D.C.Cir.1974,
. We view this requirement as more stringent than the requirement for sniffing of property, for the search is more intrusive. The requirement stated here is the “reasonable cause” standard of Terry.
. Tex.Educ. Code Ann. § 21.032 (Vernon Supp. 1982). Given the public interest in encouraging noncompulsory secondary educаtion and the social pressures to remain in school, as well as the difficulty of applying different standards to the 17- and 18-year-olds in the public school system, our discussion applies equally to those students not legally compelled to attend school.
. Terry v. Ohio, 1968,
. Courts usually refer to the basis of these powers as the in loco parentis doctrine. Under that doctrine, parents were viewed as ceding their parental powers over and duties to protect the best interests of a child to the school official, who then could act toward the child in any way that the parent could, exercising the same disciplinary and supervisory powers. Although the courts no longer view this doctrine as making the official effectively a private party, they still explain the broad powers of the school official as derived from his in loco parentis duties. We cannot accept this view. The law recognizes broad powers in a parent in part because one can safely assume that the parent will exercise those powers in the best interests of the child. The school administrator’s duties, however, are not always exercised with only the child who is being disciplined or searched in mind. On the contrary, the school official must bear in mind the interests of all the students committed to his supervision and frequently actions toward one child will be taken to protect other children from him. To that extent, it no longer makes sense to view the school official as the equivalent of a parent. See generally State v. Mercer, Tex.Civ.App.— Austin 1970,
. We intimate no opinion as to the standards to be applied when a school official acts at the request of the police, calls in the police before searching, or turns over the fruits of his search to the police. In that situation, when there is some component of law enforcement activity in the school official’s actions, the considerations may be critically different. See, e.g., Picha v. Wielgos, N.D.Ill.1976,
. Contra, Stem v. New Haven Community Schools, E.D.Mich.1981,
. We recognize that this statement leaves the fourth amendment a sliding scale in the public schools. Although а sliding scale approach has definite advantages in permitting the accommodation of all the interests in conflict in any situation, it does “convert[ ] the fourth amendment into one immense Rohrschach blot”, leaving government officials and the courts with little guidance and tempting courts to defer to discretionary exercises of power by school officials. Amsterdam, Perspectives on the Fourth Amendment, 58 U.Minn.L.Rev. 349, 393 (1974); see generally id. at 388-95. But these waters are not uncharted, and we think it too late to re-map them. The courts have established that reasonableness in the schools is to be a function of all the circumstances, including the seriousness of the problem, the age of the child, the child’s disciplinary history, and the intrusiveness of the search. See, e.g., M.M. v. Anker, 2 Cir. 1979,
. Cf. Amsterdam, Perspectives on the Fourth Amendment, 58 U.Minn.L.Rev. 349, 391 (1974) (sliding scale approach might permit warrant-less intrusions in some cases).
. Our holding is not be read broadly. More intrusive searches, and those presenting greater possibilities of abuse, will be subject to the warrant requirement. Accord, M.M. v. Anker, E.D.N.Y.1979,
. It will not, however, be enough to show that the dogs are reasonably reliable in indicating the presence or recent presence of contraband. If the reaction is to justify a search, it must give rise to reasonable suspicion that the search will produce something — i.e., reasonable suspicion that contraband is currently present.
. According to the representative of SAI, one student was scratched slightly when he played with one of the dogs, but he did not register any complaint. One who chooses to play with a large dog assumes the risk of incurring minor scratches and bruises in the rough-and-tumble, and we think that the students can differentiate between an unprovoked attack and an incidental scratch. It is unlikely that such an incident contributes to the intimidation alleged by the plaintiffs.
. We reiterate that we leave undecided whether the use of the dogs to sniff the students themselves would be a deprivation of liberty or property without due process.
. We have found no authority to the contrary, and the plaintiffs have cited none. The only cases cited by the plaintiffs deal with the possible existence of a liberty or property interest in the renewal of government employment and the possible existence of a liberty interest in one’s reputation. Bd. of Regents v. Roth, 1972,
. The judge may have thought that the plaintiffs simply failed to meet their burden of proof. As our discussion will show, we disagree.
. Technically, only the requirements of section (a) of rule 23 are “prerequisites” to a class action, and section (b) describes the categories of actions maintainable as class actions. The discussion here will treat qualification for a rule 23(b) category as a prerequisite to a class action.
. We reject the suggestion in the defendant’s brief that a putative representative must present proof of financial resources in order to meet his burden of proof on the issue of adequacy. The case relied upon by the defendant holds only that the class opponent may undertake discovery of the putative representative’s financial condition. See Klein v. Miller, N.D. Tex.1978,
. Intraclass antagonism may be analyzed under either rule 23(a)(4), the adequacy requirement, or rule 23(a)(3), the typicality requirement. See 7 C. Wright & A. Miller, Federal Practice and Procedure §§ 1768, 1769 (1972). The requirements are closely related, for demanding typicality on the part of the representative helps ensure his adequacy as a representative. We prefer to analyze the question of intraclass antagonism under the requirement that the representative protect adequately the interests of the class rather than under the requirement that his claims be typical, because each class member has the claim asserted by the plaintiffs, so the plaintiffs’ claims are typical, but many members do not see it as in their best interests to assert that claim. The real question then is whether, in spite of the typicality of their claims, the named plaintiffs can adequately represent the interests of the class, including any interest in not asserting claims.
. See, e.g., Aiken v. Nieman-Marcus, N.D.Tex. 1977,
. Hansberry, of course, does not directly forbid class actions; it simply held that some class actions will not bind the class. But rule 23(a) is designed to permit certification only in those cases in which the class can be bound and in which courts in subsequent actions would hold, consonant with due process, that res judicata barred any further litigation by class members on the cause of action. See Fed.R.Civ.P. 23(c)(3) advisory committee note. Thus Hans-berry indirectly prohibits class actions that will not bind the class.
. Fed.R.Civ.P. 23(c)(1).
. Cf. Snyder v. Bd. of Trustees of Univ. of Illinois, N.D.Ill.1968,
. A second consideration supporting our decision is that the absеntee members will not be much better protected from the effect of the decision if we deny certification. In a case like this one, the stare decisis effect of our decision that the sniffing procedures are unconstitutional will, as a practical matter, put an end to all searches. See Ihrke v. Northern States Power Co., 8 Cir. 1972,
. There may also be disagreement among class members over appropriate relief. On that issue, we cannot depend on the defendant to represent the views of all absentees. As a result, we direct certification on the issue of liability only, a procedure explicitly provided by rule 23(c)(4)(A).
. The defendant suggested in the district court but has not argued before us that the case was moot as to Robby Horton and Sandra Sanchez, who were seniors when the complaint was filed. Robby did not graduate before the district court decision, but presumably he and Sandra have both graduated, and Heather is currently a junior. Even if the certification in this case does not “relate back” to the filing of the complaint, Sosna v. Iowa, 1975,
