This is an interlocutory appeal of a circuit court’s refusal to certify a class action. The appellants are Charles and Wilma Summons who brought the action on behalf of themselves and their children, Kellen and Tanya Summons. They seek to represent a class of several thousand persons who were evacuated from their homes or businesses as the result of a railroad accident in which a chemical tank car overturned in North Little Rock. The appellees are Missouri Pacific Railroad (MOPAC), which operated the train, and Union *118Carbide Corporation, the shipper of the chemical. We agree with the Summonses’ argument that the Trial Court abused its discretion in refusing to certify the class, and thus the decision is reversed and remanded.
The accident occurred mid-morning on July 8, 1987. A liquid substance was observed to be leaking from the overturned car which was carrying ethylene oxide, an allegedly highly volatile and toxic chemical. The evacuation began with an order by local emergency services personnel at approximately 10:40 a.m. The order was lifted at approximately 4:15 p.m. when it was determined that the liquid leaking from the car was not ethylene oxide, but a non-hazardous refrigerant which was part of the overturned car’s container system.
The Summonses alleged that they and others who were evacuated were frightened, their lives were disrupted, they were forced to spend money for food, clothing, and shelter, and that they suffered pain and mental anguish. They alleged that some persons were forced to seek medical treatment. They contended that MOPAC and Union Carbide were wilfully and wantonly negligent, and they asserted a strict liability claim based on shipment of an ultra-hazardous product.
The complaint stated that the class the Summonses sought to represent consisted of about 5,000 persons, that joinder of so many claims would be impractical, that there were questions of law and fact common to the claims, and that the interests of the class outweighed those of individual members. It was also alleged that the Summonses’ claim was typical of members of the class, they had obtained competent counsel, they were aware of their responsibilities as members of the class, and a class action was the superior method of deciding the claims of the class members.
In its answer, MOPAC denied negligence, denied that the transportation of ethylene oxide was an ultra-hazardous activity giving rise to strict liability, and denied all of the allegations supporting the certification of the class. Union Carbide filed a similar answer, adding that the injuries claimed were the result of actions of parties over whom Union Carbide had no control or were the result of intervening causes.
A hearing was held on the Summonses’ motion to certify the *119class. Janet Jones, an employee of McHenry Law Firm, testified that she compiled a list of persons claiming damages as a result of the evacuation. The list numbered 5,321 persons, many of them living in the Eastgate Terrace, Shorter Gardens, and Dixie Addition sections of North Little Rock. On cross-examination Ms. Jones testified about the differences in the claims. She said some were for motel and meal expenses, and some were for loss of work because, for example, of inability of claimants to return to their homes to obtain uniforms to wear. Also, on cross-examination, Ms. Jones pointed out that she had no first-hand knowledge that anyone from MOPAC or Union Carbide assisted the North Little Rock police in the evacuation procedure.
Testimony was also taken from potential class members about the general effect of the evacuation upon their lives on the day in question.
The operative portion of the Court’s order was as follows:
1. A class action proceeding would result in indeterminate and chaotic litigation and would cause judicial extravagance rather than judicial economy.
2. A class action would be an inappropriate method of dealing with the proposed claims.
3. This action should not be maintained as a class action because the prerequisites of Rule 23 . . . have not been satisfied. ...
1. Rule 23 requirements
At the time the decision was made, Rule 23(a) contained the “prerequisites” to a class action as follows: “Where the question is one of a common or general interest of many persons, or where the parties are numerous and it is impracticable to bring all before the court within a reasonable time, one or more may sue or defend for the benefit of all.” The Rule has since been amended to list the prerequisites as follows:
(1) the class is so numerous that joinder of all members is impracticable, (2) there are questions of law or fact common to the class, (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class, and (4) the representative parties will fairly *120and adequately protect the interests of the class.
Whichever version is applied, we find the prerequisites satisfied.
a. Numerosity
The Trial Court stated in his conclusory remarks that “you could have conceivably some 3,500 plaintiffs in a case,” and that he knew of no rule which would prohibit it. There was no explanation why a class action would not be superior where there is such an “unwieldy” number of plaintiffs. Joinder of so many claims is obviously, in the language of the Rule, “impracticable.”
In City of North Little Rock v. Vogelgesang, 273 Ark. 390, 619 S.W.2d 652 (1981), we held that 17 potential plaintiffs was too small a number to satisfy the Rule. In Cooper Communities, Inc. v. Sarver, 288 Ark. 6, 701 S.W.2d 364 (1986), we held that 184 were enough. In International Union of Electrical, Radio & Machine Workers v. Hudson, 295 Ark. 107, 747 S.W.2d 81 (1988), we held that “at least several hundred” class members were enough. The fact that there are several thousand claimants in this case is enough to satisfy the numerosity requirement.
b. Common questions
MOPAC’s brief seems to concede that there are common questions of law and fact relating to its conduct and that of Union Carbide. It argues, however, that they do not predominate the issues of causation and damages which will be different with respect to each claimant. Union Carbide’s brief adds the contention that plaintiffs who suffered physical injury may be entitled to recover for mental anguish but that the many who do not claim to have suffered physical injury will have to show wilful and wanton misconduct to recover.
It may indeed be true that the class can be and will be divided into subclasses with respect to theory of recovery if it is determined that MOPAC and Union Carbide were engaged in mere negligence, wilful and wanton negligence, or an ultrahazardous activity. The predominate question, however, in any case will be factual. The defendants’ actions will have to be determined, and then it will have to be determined if those actions were wrongful and to what degree. We have no hesitancy in *121saying there are common questions.
As to whether the common questions predominate others, such as causation in terms of class definition and damages, there is a clear overlap with the question of whether the class action is a superior method of handling the litigation in prospect. See H. Newberg, Class Actions, § 4.22 (1985). We will discuss it below along with the abuse of discretion standard of review.
c. Typicality
MOPAC and Union Carbide contend that the Summonses’ claim or claims are not typical of those of the other members of the class they seek to represent because, unlike some members, the Summonses allege no physical harm and only seek to recover for their inconvenience and fear in having to leave their home and for the expense of eating out.
The typicality requirement is discussed in H. Newberg, Class Actions, § 3.13, supra, as follows:
Typicality determines whether a sufficient relationship exists between the injury to the named plaintiff and the conduct affecting the class, so that the court may properly attribute a collective nature to the challenged conduct. In other words, when such a relationship is shown, a plaintiff’s injury arises from or is directly related to a wrong to a class, and that wrong includes the wrong to the plaintiff. Thus, a plaintiff’s claim is typical if it arises from the same event or practice or course of conduct that gives rise to the claims of other class members, and if his or her claims are based on the same legal theory. When it is alleged that the same unlawful conduct was directed at or affected both the named plaintiff and the class sought to be represented, the typicality requirement is usually met irrespective of varying fact patterns which underlie individual claims. [Footnotes omitted.]
Although the Summonses’ allegations as to their injuries and damages are different from those they describe for other members of the class, their claims are typical in the sense that they arise from the alleged wrong to the class which includes the wrong allegedly done to them, and that is sufficient.
*122 d. Representation
Mrs. Summons testified that she understood her obligations in undertaking representation of the class and the possible costs involved. She said she would do whatever was necessary in that respect. The Summonses have alleged that they are represented by counsel competent to handle a class action, and MOPAC and Union Carbide have given us no reason to doubt that allegation.
2. Rule 23 (b)
Whether common questions of law or fact predominate and whether a class action is a superior method of deciding the case are, to a degree, necessarily subjective questions and very much related to the broad discretion conferred on a trial court faced with them.
In International Union of Electrical, Radio and Machine Workers v. Hudson, supra, we discussed the line of cases in which we have held that a trial court has broad discretion to protect the interests of the members of the class and to allow or disallow an action to proceed as a class action. The essence of the case now before us is the decision we must make whether the Trial Court abused its discretion in refusing the certification, and the questions of predominance and superiority are integral to that decision.
If the predominance of the liability aspects of the claims of the class members were to be judged solely on the basis of the number of cases in which the same issues of liability are to be decided, predominance of the common questions would undoubtedly be established. We recognize, however, that there are other factors to be considered.
In its brief, MOPAC emphasizes its contention that each of the potential class members will have to demonstrate the element of causation, and cites Ryan v. Eli Lilly & Co., 84 F.R.D. 230 (D.S.C. 1979), in which a United States District Court determined a class action would be inappropriate. The claim in that case was on behalf of daughters of mothers who had been administered a drug manufactured by the defendant. The drug which was used by the mothers resulted in an abnormal vaginal *123condition in the daughters. The Court discussed the fact that there would be varying factors with respect to liability as follows:
In deciding the issue of predominance this Court must predict the evidence likely to be introduced at trial. This is particularly true when purported class members are not standardized. See Windham v. American Brands, Inc., 68 F.R.D. 641 (D.S.C. 1975). Standardized in this sense refers to a class of individuals whose claims against a common defendant arise out of essentially identical fact patterns. The case at bar presents a different class of plaintiffs. The mothers of the proposed plaintiffs in this case each used a synthetic estrogen; however, the length of exposure, the reason for the drug’s use, the specific chemical formulation of the drug, the state of the art at the time of consumption or the manufacturer’s knowledge of synthetic estrogen’s carcinogenic effect and possible medical result in the absence of the estrogens are all specific points going toward proximate causation which will require proof for each individual class member. The Court is aware of decisions such as Samuel v. University of Pittsburgh, 506 F.2d 355 (3rdCir. 1974); wherein a bifurcation of trials between the liability and damages issues was held appropriate. Bifurcation here would not cure the deficiencies of the instant class action, since the liability issue alone will require individual proof as to each party plaintiff. It is the necessity of such separate evidentiary showings that leads this Court to conclude the common questions of fact do not predominate.
In this case, the liability question may be complicated in terms of the theories of liability advanced, however, those issues may be the same with respect to each of the claimants, and unlike the Ryan case, they will at least be the same as to members of subclasses such as those who claim physical injury and those who do not.
Cases such as Strauss v. Long Island Sports, Inc., 60 A.D. 501, 401 N.Y.S.2d 233 (1978), cited by Union Carbide, are not as easily distinguished. In the Strauss case, certification of a class of plaintiffs claiming injury from misleading advertising was held to be an abuse of discretion because each member of the class *124would have had to prove reliance. While we might say that that holding differs because each member of the class would have to prove an element of the tort, we cannot see a real difference from this case in which each claimant must, after proving basic negligence or other tortious conduct, prove causation, or that he or she was uprooted, disrupted, and otherwise injured by the acts of the defendants. We can only assume that the Strauss case is a traditionally hostile interpretation of the New York class action rule of the sort from which we are willing to part.
The Court in the Ryan case points out that the Advisory Committee to the Federal Rules of Civil Procedure commented that the class action is “ordinarily not appropriate” in “mass accident” torts because questions not only of damages but of liability and defenses to liability may affect different claimants in different ways. In 7B C. Wright, A. Miller, and M. Kane, Federal Practice and Procedure, § 1783 (1966), both sides of the controversy are presented. The other side is as follows:
In many ways [Federal] Rule 23(b)(3) seems particularly appropriate for some tort cases of this type. Thus, some courts have ignored the Advisory Committee’s doubts and have allowed mass tort cases to be brought under subdivision (b)(3). The central issue of liability, for example, may be a difficult one that occasionally will require lengthy expert testimony, perhaps concerning the physical condition of a vehicle or the state of a technological art in a particular field of transportation or manufacturing. If the various tort claims were tried individually, the evidence would have to be repeated time and again. . . . [Footnotes omitted.]
The repeated litigation of the liability question and the attendant possibilities of inconsistent results in a case like this one outweigh the admitted fact that each claimant will have different damages evidence. A factor we recognized in Drew v. 1st Fed’l S & L Ass’n of Ft. Smith, 271 Ark. 667, 610 S.W.2d 876 (1981), and in the International Union case, and which is central to the result we reach here is the basic consideration of fairness. An ingredient of that consideration is the point of our decision in the International Union case:
By limiting the issue to be tried in a representative *125fashion to the one that is common to all, the trial court can achieve real efficiency. The common question here is whether the unions can be held liable for . the actions of their members during the strike. If that question is answered in the negative, then the case is over except for the claims against the named individual defendants which could not be certified as a class action. If the question is answered affirmatively, then the trial court will surely have ‘splintered’ cases to try with respect to the damages asserted by each member of each of the subclasses, but efficiency will still be achieved, as none of the plaintiff's would have to prove the unions’ basic liability.
Is that unfair? It is not unfair to the unions, as they will be able to defend fully on the basic liability claim, and they will have the opportunity to present individual defenses to the claims of individual class members if their liability has been established in the first phase of the trial. They lose nothing. Would it be fair to the class members to require them to sue individually? The evidence so far shows that each putative class member has a claim that is too small to permit pursuing it economically. If they cannot sue as a class, the chances are they will not sue at all. We agree with the unions’ argument that the sole fact that the claims are small is not a reason to permit a class action, but it is a consideration which has appeared when other courts, as we must do, have considered whether the class action is superior to other forms of relief. See C. Wright, A. Miller, and M. Kane, supra, § 1779, n. 21, citing Roper v. Consurve, Inc., 578 F.2d 1106 (5th Cir. 1978), affirmed on other grounds, sub nom. Deposit Guar. Nat’l Bank v. Roper, 445 U.S. 326 (1980); Werfel v. Kramarsky, 61 F.R.D. 674 (D.C.N.Y. 1974); and Buchholtz v. Swift & Co., 62 F.R.D. 581 (D.C. Minn. 1973).
We recognize that the trial court has substantial power to manage a class action even though the directions given in our Rule 23 are not as extensive as those given in the comparable federal rule. This power to manage the action contributes to the discretion we find in the trial court to determine whether a class should be certified. We conclude there was no abuse in this case.
*126Our opinion in the International Union case was hardly an exhaustive review of all the considerations leading to the conclusion that the class action is the superior manner of deciding the typical mass tort case. That has been done, however. D. Rosenberg, Class actions for Mass Torts: Doing Individual Justice by Collective Means, 62 Ind. L. Rev. 560 (1986-1987). Here are some excerpts from the article:
In mass tort cases involving claims or personal injury, which pose daunting problems of causation and remedy, the price of individual justice is notoriously high. Because they typically involve complex factual and legal questions, mass tort claims are exceedingly, if not prohibitively, expensive to litigate. The questions of whether the defendant’s conduct failed to satisfy the governing standard of liability frequently entail interrelated technological and policy issues that require extensive discovery, expertise, and preparation to present and resolve adequately. Equally demanding are the causation issues in mass tort cases, such as whether the plaintiffs condition was caused by exposure to the substance in question or to some other source of the disease risk.
The case-by-case mode of adjudication magnifies this burden by requiring the parties and courts to reinvent the wheel for each claim. The merits of each case are determined de novo even though the major liability issues are common to every claim arising from the mass tort accident, and even though they may have been previously determined several times by full and fair trials. These costs exclude many mass tort victims from the system and sharply reduce the recovery for those who gain access. Win or lose, the system’s private law process exacts a punishing surcharge from defendant firms as well as plaintiffs.
***
These conditions generally disadvantage claimants. Because defendant firms are in a position to spread the litigation costs over the entire class of mass accident claims, while plaintiffs, being deprived of the economies of scale afforded by class actions, can not, the result will usually be that the firms will escape the full loss they have *127caused and, after deducting their attorney’s shares, the victims will receive a relatively small proportion of any recovery as compensation. As a consequence, the tort system’s primary objectives of compensation and deterrence are seriously jeopardized.
***
Because of their cost-spreading advantages, a defendant firm typically can afford not only to invest more in developing the merits of the claim than the opposing plaintiff attorney, but also to finance a “war of attrition” through costly discovery and motion practice that depletes the adversary’s litigation resources. The consequences of redundantly litigating common questions thus skews the presentation of the merits, promotes abusive strategic use of procedure, needlessly consumes public resources, and ultimately drains away a large amount of the funds available to redress by judgment or settlement, victim losses.
We can cite no case like this one where it has been held that the trial court abused its discretion by denying certification. We can, however, cite a very similar case where it was held that the trial court did not abuse its discretion in certifying a class alleging injury as the result of a chemical fire caused by a railroad accident. Reynolds v. CSX Transportation, Inc., 55 Ohio App.3d 19, 561 N.E.2d 1047 (Ohio App. 1989). In that case, a master recommended against certification because there would be five ill-defined sub-classes of plaintiffs. The Trial Court disagreed. The Court of Appeals held there was no abuse of the Trial Court’s discretion.
In applying Ohio Civil Procedure Rule 23A, which was apparently like the Federal Rule and like the current Arkansas Rule, and Rule 23B containing the same rather subjective criteria as the parallel sub-sections of the Arkansas and Federal Rules, the Trial Court was quoted as follows:
One time answers can be had to the issues of negligence and malice, such as: What were the duties of the respective defendants? What did they undertake to do or not do in respect to those duties, if any? Was there a breach *128of duty? Was there a conscious disregard for the rights and safety of others that had a great probability of causing substantial harm?
The Court of Appeals then wrote:
We conclude that the trial court’s decision to certify the plaintiffs’ case as a class action to the issues of negligence and malice was not an abuse of discretion. The court’s order was not unreasonable, arbitrary or unconscionable. Rather, it was a well-reasoned decision based upon the trial court’s unique ability to determine which procedural devices are best for handling such unwieldy cases.
As noted above, the Reynolds decision was a bow to a trial court’s discretion, and thus hardly in direct support of our conclusion here. It does, however, provide an example of a case where that which the plaintiffs ask here has been found to be reasonable under very similar circumstances.
We are well aware of our remark in the International Union case that, had the Trial Court refused to certify the class we might well have affirmed, but we have not held that a trial court’s discretion is so broad on this issue that it cannot be reviewed. If a trial court’s decision not to certify a class action in this type case is at all reviewable, then this is the time and the sort of case in which to review it. That is especially so when it is possible that a large number of persons who may have legitimate claims not worth pursuing because of the costs of our system of justice may lose those claims if they are not allowed to proceed together as a class. By not certifying a class, a trial court can cause the problem to “go away” to the extreme disadvantage of the claimants unless that decision is reviewable. It is especially important that we review an order such as the one in this case given the newness of our decision in the International Union case and the recent changes in Rule 23(a).
Given our conclusion that the Trial Court erred in stating that the facial prerequisites of Rule 23(a) were not satisfied, we are left with consideration of the Trial Court’s remarks at the conclusion of the certification hearing and the final order. Most of the factors he noted seem to us to support *129certification. The language of the Trial Court’s order is not helpful to an understanding of the decision because it consisted only of conclusions without reasons.
We conclude this opinion with a caveat; the only issue we have addressed is whether the class should have been certified. We have not intimated, nor have we meant to intimate, that any of the class members will or should recover. The question of liability is not at all before us.
The order refusing to certify the class is reversed and remanded for proceedings consistent with this opinion.
Hays, J., dissents.
Brown, J., concurs.