A jury found Phillip Coleman guilty of four counts of being a felon in possession of a firearm, see 18 U.S.C. Sec. 922(g)(1) (as well as one count of unlawful possession of an unregistered firearm, see 26 U.S.C. Sec. 5861(d); a conviction not in issue here). Coleman attacks the gun convictions on three grounds: 1) that the four felon-in-possession counts were misjoined (or, in the alternative, erroneously not severed); 2) that for one of those counts the government's evidence was insufficient to establish that the gun traveled in interstate commerce--a necessary element of the offense; and 3) that for the same count the underlying search and seizure was unreasonable. We are not persuaded by any of these submissions and accordingly affirm all of the convictions.

The four possession counts were premised upon four separate incidents occurring on April 18, 1990, January 14, 1992, May 23, 1992, and June 2, 1992. While the first two are essentially unconnected with each other and with the latter two, there is slightly more than mere temporal proximity linking the events of May 23 and June 2. On May 23, Milwaukee police officers observed Coleman pointing a small gun at another man and gave chase on foot. After catching up with Coleman and recovering a small gun from underneath a dumpster located along the route Coleman took, the officers ran a warrant check on Coleman and found none outstanding. Coleman told the officers that he was chasing three men who had tried to break into his house, and because of a mixup among the officers, Coleman was released. Five days later one of the officers telephoned Coleman's parole officer, Susan Dybul, and told her that Coleman was seen chasing another man with a handgun. (Coleman was on parole after serving prison time for weapon possession, burglary and attempted murder; possession of firearms violated the conditions of his parole.) The officer also told Dybul that he was planning to refer the case to federal authorities. Later that day, Coleman called Dybul and reported that he was arrested and released the previous day in a separate incident--driving with a revoked license. Planning to take him into custody because of the weapon possession report, Dybul told Coleman to report to her office on June 2 at 9:30 a.m. Meanwhile, a federal agent, who had been contacted by the Milwaukee police, called Dybul and arranged to arrest Coleman on federal charges when Coleman appeared at Dybul's office.

On June 2, before Coleman was due to arrive, Dybul received a call from Cheryl Williams who claimed that Coleman had stolen her pistol from her purse six days earlier because she owed him seventy dollars for cocaine he had provided. In the conversation, Williams identified herself, provided an address and two telephone numbers, described the gun and confirmed that it was bought and registered in her name. Dybul told Williams to file a police report, which she did. Because of Williams' call, Dybul then met with her supervisor, and they discussed conducting a search of Coleman's home. (Under applicable procedures, Dybul and her supervisor would need the approval of their assistant regional chief to conduct a home search.) A short time later, the federal agents arrived as arranged, and Dybul informed them about Williams' call and the parole search which was being planned as a result. She asked them, and they agreed, to assist in the search by providing backup. In the meantime, Dybul's supervisor received permission to carry out the search, and the Milwaukee police phoned Dybul to inform her that Williams had filed a report regarding her missing gun. When Coleman arrived at Dybul's office at 9:30, the federal authorities arrested him. Soon afterwards, Dybul, her supervisor and two federal agents drove to Coleman's residence and entered the home using a key that was seized from Coleman. After the federal agents conducted a brief protective sweep, Dybul and her supervisor searched the house. On the second floor, Dybul found a sawed-off shotgun which was to become the subject of Count III of the indictment.

Coleman argues that the shotgun should have been suppressed as a fruit of an illegal search. Acknowledging that Wisconsin's procedure for warrantless searches of probationers' homes is constitutional,1 see Griffin v. Wisconsin, 483 U.S. 868, 107 S. Ct. 3164, 97 L. Ed. 2d 709 (1987), Coleman asserts that federal and state authorities colluded to circumvent the warrant requirement which would normally bind a federal investigative search of his home by jointly conducting the June 2 search and pretextually labeling it a probation search. In the alternative, Coleman maintains that there were not "reasonable grounds" to believe that his home contained contraband as required by the Wisconsin procedure. See Wisconsin Administrative Code DOC Sec. 328.21(a).

These arguments we can reject in short order. First of all, while we agree that federal law enforcement officers (or the police in general) cannot utilize state probation officials to carry out warrantless searches on their behalf which they as federal agents, acting alone, could not execute without a judicial warrant, see, e.g., Shea v. Smith, 966 F.2d 127, 132 (3d Cir.1992); United States v. Harper, 928 F.2d 894, 897 (9th Cir.1991); United States v. Cardona, 903 F.2d 60, 65 (1st Cir.1990), cert. denied, 498 U.S. 1049, 111 S. Ct. 758, 112 L. Ed. 2d 778 (1991), there is no basis upon which to conclude that anything of the sort occurred here. The district court found, and all testimony supports, that Dybul and her supervisors decided on their own to search Coleman's residence as a result of the tip they received from Cheryl Williams. Simply nothing except sheer speculation indicates otherwise. That final approval for the parole search was not finally obtained from the assistant regional chief until the federal agents had arrived at Dybul's office in order to arrest Coleman (as prearranged) by itself proves nothing and certainly does not render the district court's findings in this matter, which were in accordance with the testimony of all relevant parties, clearly erroneous. See United States v. Dent, 984 F.2d 1453, 1459 (7th Cir.), cert. denied --- U.S. ----, --- U.S. ----, 114 S. Ct. 169, 114 S. Ct. 209, 126 L. Ed. 2d 129, 126 L. Ed. 2d 165 (1993). And the state parole officials were entitled to seek the assistance of handy federal agents to ensure a safe search. Cf. Harper, 928 F.2d at 896 (stating that a "parole officer [can] cooperate[ ] with the police to achieve her own legitimate objectives"); Cardona, 903 F.2d at 66 (concluding that "police officers and parole officers are fungible when the former serve as mere implementers decisions already made by the latter").

Coleman's alternative claim that Dybul did not have "reasonable grounds" to believe his residence contained contraband as required under Wisconsin's probation search procedure is also without merit. In Griffin, the Supreme Court found that the Wisconsin procedure, with its definition of "reasonable grounds," comports with the Fourth Amendment's reasonableness requirement and thus a search carried out pursuant to it is valid. Accordingly, if Dybul's search of Coleman's home complied with that scheme, the search passes constitutional muster.2 "Reasonable grounds" denotes a degree of certainty less than that of probable cause, see Wisconsin v. Griffin, 131 Wis. 2d 41, 388 N.W.2d 535, 539-44 (Wisc.1986), aff'd 483 U.S. 868, 107 S. Ct. 3164, 97 L. Ed. 2d 709 (1987), and is informed by a host of factors,3 of which a number are clearly relevant here and show the propriety of the search in this case. In short: When Cheryl Williams made her call, Dybul already had a police report that Coleman was recently seen possessing a handgun, which indicated, in the district court's words, his "recent predisposition to carry handguns." Charged with supervising parolees, Dybul could (and should) have felt the especial need to ensure compliance with this particular condition of Coleman's parole, as violent felons and firearms are often a poor mix. Moreover, Williams' information was both detailed and accompanied by indicia of reliability. Williams provided her name and where she could be contacted and when prompted repeated the information to the police in a formal report. While Dybul knew that Williams owed Coleman money (indeed, for drugs), the fact that Williams would disclose this information in such an upfront way could reasonably reassure Dybul that Williams was being truthful. Perhaps if more were known about Williams and her relationship with Coleman, her credibility should have been discounted (or possibly enhanced). But looking solely at the information before Dybul at the time the decision was made to search Coleman's home, we are in complete agreement with the district court's conclusion that the search was conducted in compliance with Wisconsin's "reasonable grounds" standard. Coleman has not cited nor could we find a Wisconsin case offering an interpretation or construction of the standard that would indicate otherwise. The district court's denial of his suppression motion was not clearly erroneous.

Coleman argues that his conviction for possession of the shotgun found by Dybul should be reversed anyway because evidence sufficient to prove beyond a reasonable doubt that the weapon moved in interstate commerce was not presented to the jury. If true, the government concedes, such an oversight would constitute a defect mandating reversal. See United States v. Buggs, 904 F.2d 1070, 1076 (7th Cir.1990). The requisite interstate commerce nexus can be established, however, merely by evidence that the gun was manufactured outside the state in which it was possessed (which everyone in this case agrees is Wisconsin). See id. As Coleman points out, the ATF expert witness who did testify particularly about the out-of-state manufacture sites of the other weapons did not specifically indicate where the Savage Arms 12 gauge shotgun subject of Count III was manufactured. To a query whether any of the four guns in the case were produced in Wisconsin, the agent did seem to respond in the negative, but his answer was ambiguous enough in its reference to all the guns that we would hesitate to hold that this testimony alone was sufficient to establish foreign manufacture of the shotgun.4 However, the agent also testified that guns are labeled with manufacturer proof marks which identify a weapon's place of manufacture, and the shotgun (which was in evidence) was in fact prominently stamped:

CHICOPEE FALLS, MASS U.S.A.

In closing argument, the government explicitly invited the jury to look at the markings on the guns for proof of interstate commerce nexus, while Coleman, despite the testimony of the ATF expert, urged the jury to consider that the markings may only represent the home office of the manufacturing corporation and not the actual place of manufacture. The jury was entitled to inspect the weapon for themselves and draw the conclusion, based on the ATF agent's testimony, that the markings were dispositive of out-of-state production. The record is certainly not "barren of any evidence, regardless of weight, from which the trier of fact could find [this element] beyond a reasonable doubt." United States v. Kaufmann, 985 F.2d 884, 895 (7th Cir.) (citations omitted), cert. denied --- U.S. ----, 113 S. Ct. 2350, 124 L. Ed. 2d 259 (1993).

Coleman contends that the four separate weapons possession charges were misjoined or at least should have been severed. Joinder of offenses is proper when, on among other occasions, the offenses "are of the same or similar character." Fed.R.Crim.P. 8(a).5 Although Rule 8(a) speaks to nothing more than the similarity between joined counts, this court, following the lead of the eighth circuit, has opined that this kind of joinder is permitted if "the counts refer to the same type of offenses occurring over a relatively short period of time, and the evidence as to each count overlaps." United States v. Koen, 982 F.2d 1101, 1111 (7th Cir.1992) (quoting United States v. Shue, 766 F.2d 1122, 1134 (7th Cir.1985) (quoting United States v. Rodgers, 732 F.2d 625, 629 (8th Cir.1984))); see also United States v. Quintanilla, 2 F.3d 1469, 1482 (7th Cir.1993); United States v. L'Allier, 838 F.2d 234, 240 (7th Cir.1988). If this is indeed the exclusive standard against which to measure the propriety of joinder of "same or similar" offenses, Coleman has a point when he suggests that joinder in this case strains the limits of a reasonable reading of that language and thus is improper. While the offenses involved here are identical in type, their temporal and evidentiary connections with one another range from moderate to quite slim. Most glaring is the government's decision to join a possession charge from 1990 with three from 1992. While the later incidents occurred within four and a half months, the earliest one dates back another twenty-one months.6 Furthermore, the incidents in question are essentially discrete, unconnected events.7 They comprise four separate encounters with the police under four different circumstances and involve four separate sets of witnesses and evidence. The government does not contest Coleman's submission that "[t]here is absolutely nothing about these counts which in any way make them relevant to any fact in issue in any other count," except for the fact that they all rely on the same predicate felony to which Coleman, at the beginning of trial, refused to stipulate (and which the government, as a result, had to prove up with the brief testimony of three witnesses). What the actual burdens at trial turned out to be, however, is irrelevant to the joinder (as opposed to severance) inquiry. "The test of misjoinder is what the indictment charges, not what the proof at trial shows." United States v. Quintanilla, 2 F.3d 1469, 1482 (7th Cir.1993). And, in any event, the focus of a Sec. 922(g)(1) case that reaches the jury is usually--and was here--the issue of possession, not whether the defendant does in fact have the status of convicted felon. It is clear that the evidentiary overlap between the joined offenses in this case was as scant as the overall temporal proximity between them was slight.

The short-period-of-time/evidence-overlap formula reflects a concern that the liberal joinder rule not result in the unchecked bundling of offenses. It recognizes that the value of joining offenses in a particular case depends upon the extent to which real efficiencies can be realized with minimal concomitant prejudice to the conduct of a fair trial. Judicial economy and convenience are the chief virtues of joint trials--i.e. joinder often avoids expensive and duplicative multiple trials--see Archer, 843 F.2d at 1021; United States v. L'Allier, 838 F.2d 234, 240 (7th Cir.1988), while defendant embarrassment or confoundment in presenting separate defenses simultaneously, jury cumulation of evidence, and jury inference of criminal disposition are its main vices. See Drew v. United States, 331 F.2d 85, 88 (D.C.Cir.1964). Courts have approximated the potential for such benefits and costs in particular cases by relying on shorthand formulations similar to the one which we have approved. Evidentiary overlap, most notably, strongly informs both sides of the equation. Separate counts that involve much of the same evidence would, if tried separately, engender duplicative efforts at trial. On the other hand, separate counts that for the most part depend on separate evidence save fewer steps when tried together. While only one jury need be impanelled and one trial judge's calendar arranged and background information elicited only once, no comparable savings are realized with respect to the substantive portions of the trial; basically the same complement of witnesses, documents and exhibits will be mustered a single time, whether in one courtroom or two. Similarly, to the extent evidence of one offense would be independently admissible at a separate trial for a joined offense, the additional prejudice a joint trial theoretically could produce is reduced. If, by contrast, evidence of the joined offenses would be inadmissible at separate trials, joinder seems to implicate the set of concerns underlying the so-called propensity rule of evidence. See Fed.R.Evid. 404.8 Of course, the proxy is a rough one and thus "does not require that every piece of evidence of one offense be admissible in a separate trial on the other offense. Rather, the joinder rule looks in a broader sense to whether the rules relating to 'other crimes' evidence have been satisfied." L'Allier, 838 F.2d at 240.

But is it precisely the joinder rule (or this joinder rule) that should be informed by such a necessarily discretionary estimate of the evidence? Interestingly, the eighth circuit's formulation in which time separation and evidentiary overlap between offenses are prominent may have originated as a guide not to joinder of offenses of "same or similar character" but to joinder of offenses that form a common plan or scheme, see United States v. McKuin, 434 F.2d 391, 395-96 (8th Cir.1970) (applying standard to offenses linked as parts of a common scheme), cert. denied, 401 U.S. 911, 91 S. Ct. 875, 27 L. Ed. 2d 810 (1971); Johnson v. United States, 356 F.2d 680, 682 (8th Cir.) (devising the standard apparently--although not clearly--for such a situation), cert. denied, 385 U.S. 857, 87 S. Ct. 105, 17 L. Ed. 2d 84 (1966), or as a standard for Rule 14 discretionary severance but sloppily presented so as to appear to be based upon Rule 8(a). See id. at 681-82 (appeal and analysis in the case was based on the denial of a motion for severance, not misjoinder). More fundamentally, the considerations of the formula are not necessary to an understanding of what "offenses ... of same or similar character" means. This language in Rule 8(a) is a rather clear directive to compare the offenses charged for categorical, not evidentiary, similarities.9 Further, the similarity of character of different offenses does not significantly depend on their separation in time. Two armored car robberies committed months apart are offenses of same or similar character; possessing five kilograms of cocaine and defrauding a bank, even if they occur on the same day, are not. Simply put, if offenses are of like class, although not connected temporally or evidentially, the requisites of proper joinder should be satisfied so far as Rule 8(a) is concerned.10 This was the thrust of earlier cases applying the Rule, see, e.g., United States v. Quinn, 365 F.2d 256, 264 (7th Cir.1966); Cupo v. United States, 359 F.2d 990, 993 (D.C.Cir.1966), cert. denied, 385 U.S. 1013, 87 S. Ct. 723, 17 L. Ed. 2d 549 (1967); Drew, 331 F.2d at 87; Finnegan v. United States, 204 F.2d 105, 109 (8th Cir.) ("It is, however, contended that the counts could not properly be joined in one indictment because each charged a separate and distinct offense, unrelated to each other. However, distinct offenses may properly be joined if they are of the 'same or similar character.' The crimes.... belong to the same class, and may be separately charged in a single indictment.") (internal citations and quotations omitted), cert. denied, 346 U.S. 821, 74 S. Ct. 36, 98 L. Ed. 347 (1953), and seems to be the current approach to joinder of "same or similar" offenses in several of our sister circuits. See, e.g., United States v. Muniz, 1 F.3d 1018, 1023 (10th Cir.), cert. denied, --- U.S. ----, 114 S. Ct. 575, 126 L. Ed. 2d 474 (1993); United States v. Holloway, F.3d 307, 310-11 (5th Cir.1993). In addition, on a number of occasions we have refocused our joinder inquiry away from a balancing of time and evidence factors and toward a more literal reading of the Rule. See, e.g., United States v. Donaldson, 978 F.2d 381, 391 (7th Cir.1992); Archer, 843 F.2d at 1021 ("[B]oth counts charge Archer with possessing prohibited objects with the intent to facilitate an escape from prison. The elements to be proved in each case were the same. This similarity supports the district court's decision of joinder.").

This approach is most consistent with the respective roles of Rule 8 and Rule 14. Rule 8 joinder is based solely on the face of the indictment and is accordingly a question of law, subject to plenary appellate review. See United States v. Shue, 766 F.2d 1122, 1134 (7th Cir.1985); see generally United States v. Werner, 620 F.2d 922, 926-29 (2d Cir.1980). The initial decision to join is made (and properly reviewed) without real insight into the actual mutual relevance of the individual offenses and their surrounding circumstances, which only crystallizes with the presentation of evidence at trial. An uncomplicated inquiry and review of initial joinder is the natural product of Rule 8, and of the practicalities of litigation, and makes sense so long as authority remains for monitoring the continued appropriateness of a joint trial as proceedings go forward, further developments unfold and evidentiary and other sources of prejudice can be better observed and analyzed. Rule 14 provides such authority,11 and pursuant to it a flexible examination of the nature of the evidence and ties between the acts underlying the offenses charged seems most appropriate. As a case proceeds beyond the indictment stage, such an examination becomes less speculative and more facilitative of the trial judge's task of calculating the actual risk of unfair prejudice that a joint trial would entail. As a result, severance decisions under Rule 14 are inevitably discretionary matters best informed by the observations and experience of the overseeing judge who is also singularly situated to accurately assess the cost of separate trials. See Koen, 982 F.2d at 1112-13.

Because Rule 8 does represent a strong preference, in language and policy, for broadly construing the propriety of joinder as an initial matter, see Archer, 843 F.2d at 1021, district courts must not shirk their duties under Rule 14. They should vigilantly monitor for developing unfairness and should not hesitate to order severance at any point after indictment if the risk of real prejudice grows too large to justify whatever efficiencies a joint trial does provide. We review Rule 14 severance decisions--which are most frequently denials--for abuse of discretion out of a recognition that the district court is peculiarly qualified to discharge this function, not to signal any general disapproval of severance when appropriate in the interest of justice. Indeed, when offenses are joined because of their "same or similar character," the risk of unnecessary unfairness infiltrating the joint trial is elevated. Unlike instances of joinder of offenses based on the same or closely connected acts or transactions, "[t]here is no comparable saving of trial time when offenses that are related only by being of the same type are joined, since the offenses are usually proven by different bodies of evidence [and], when totally unrelated, similar offenses are joined, defendant faces a 'considerable risk' of prejudice." 8 Moore's Federal Practice p 8.05 (citations omitted). In such cases, the district courts should be especially watchful for possible jury confusion, illegitimate cumulation of evidence or other sources of prejudice not worth the reduced efficiency gains of a joint trial.

Here, joinder under Rule 8(a) was appropriate because Coleman was charged with four counts of possession of a firearm, identical 922(g)(1) offenses which could only vary in time and location but not in their essential elements. Furthermore, we cannot conclude under the circumstances of this case, considering especially the amount and nature of the evidence and the nature of the joined counts, that the denial of Coleman's severance motion was an abuse of the district court's discretion. "[T]he evidence as to each was short and simple; there was no reasonable ground for thinking that the jury could not keep separate what was relevant to each." Lotsch, 102 F.2d at 36. The incidents were discrete as was the proof offered at trial, and the evidence as to each was anything but excessive or confusing. Also, the central contested issue for each count was virtually the same--i.e. constructive possession--and, as a result, the jury did not have to grapple with the application of widely variant governing legal principles. Their task thus was at all times and in all respects a manageable one, and we do not believe that such social opprobrium generally attaches to accusations of gun possession to render it unlikely that the jury could fairly and conscientiously weigh and consider the evidence for each count before reaching a verdict.12 The jury was instructed to consider each count and the relating evidence separately, see Koen, 982 F.2d at 1113; there was no reason to suppose that it would disregard this mandate.

Because we conclude that the district court did not err by allowing the charges to be tried together, by denying the suppression motion or by sustaining the jury's verdict, all the challenged convictions are affirmed.