The opinion of the court was delivered by

DAVIS, J.:

The defendant, Aaron Pennington, appeals from his convictions of first-degree murder and aggravated burglary. He contends that the trial court erred (1) by admitting evidence of an experimental shooting test; (2) by failing to instruct the jury that it must not draw inferences from the defendant's failure to testify; (3) by instructing the jury on aiding and abetting; and (4) by failing to give the cautionary jury instruction on the testimony of a paid informant. Finding no reversible error, we affirm.

On July 20, 1991, David Brown discovered his mother's body in her bed when he stopped by her home to check on her. Laura Brown had been shot three times in the head with a .22 caliber weapon. The police discovered a pillow from the living room couch in a chair next to the front door. The pillow had bullet holes in it, and police surmised the killer had used it to muffle the sound of the shots. There were no signs of forced entry, and nothing of value had been taken from the house, although the victim's purse had been emptied onto the bed beside her body.

The defendant was a friend of Mrs. Brown's other son, Paul Brown. A few months before the murder, the defendant had spent several nights in Mrs. Brown's garage without her knowledge. When she discovered the defendant's presence, she required the defendant to leave. One of the defendant's friends testified that the defendant was not happy about being asked to leave.

On the night of the murder, the defendant threw a party at Motel 6 in Lenexa. He paid for the room and much of the liquor. Such generosity was uncharacteristic of the defendant. He had been unemployed and homeless for several months before the homicide.

During the party, the defendant obtained a ride from two people at the party to a Quik Trip store located about two and one-half blocks from Mrs. Brown's home. On the way to the Quik Trip he changed from his tennis shoes into work boots and put on surgical gloves. He told the others in the car that he would call them at Motel 6 to pick him up. He called later for a ride back to the motel. The defendant told the party guests that if anyone asked where he was, they were to say he was with them all night at Motel 6.

The defendant left his work boots in the trunk of the car in which he had ridden to the Quik Trip. The car owner saw a handgun inside one of the defendant's boots and told the defendant to get the gun out of his car. The defendant refused at first. A few days later the defendant got his gun out of his friend's car and threw it over a bridge into the Kansas River. A passenger in the car in which the defendant rode to the bridge testified that the gun that the defendant disposed of held nine rounds, rather than the standard six rounds.

Paul Brown owned a .22 caliber handgun with a nine-round chamber. Brown first told police that he did not own a gun. He later admitted owning a gun, but told the police that it had been stolen the week leading up to the homicide. Although Brown told police that he spent the entire night of the homicide with a friend, two neighbors testified that they saw Brown's van in front of his mother's house at 3:00 a.m. and 4:30 a.m. during the night of the murder.

(1) Experimental Test

The defendant claims that the testimony of a neighbor three houses east of the victim's home established that the fatal shots were fired at 3:30 a.m. He claims that other witnesses established that he was back at Motel 6, nowhere near the victim's home, at 3:30 a.m. The testimony of the neighbor depended upon his ability to hear the shots at 3:30 a.m. Over the defendant's objection, the State was allowed to introduce evidence of an experimental test conducted several months after the shooting. Three shots were fired under what the State claims were similar conditions, and the neighbor was unable to hear the shots fired.

The defendant argues that the conditions were not substantially similar to the conditions at the time of the homicide and that evidence of the experimental test was inadmissible. The parties do not disagree on the law but argue about its application to the facts of this case.

The admissibility of scientific or experimental tests lies within the sound discretion of the trial court. As we said in State v. Jones, 202 Kan. 31, 42, 446 P.2d 851 (1968), "It is well established that scientific experiments may be shown in evidence as well as the testimony of experts in support thereof (K.S.A. 60-456; 29 Am.Jur.2d, Evidence §§ 818, 819, 822, 824, pp. 908-912)." We acknowledged in Jones that "evidence of an occurrence entirely disconnected with the one involved which tends to illustrate a physical fact, where the conditions are the same or similar, is relevant and admissible because the observed uniformity of nature raises an inference that like cause will produce like results." 202 Kan. at 42.

The key to admissibility is relevance. As noted in our most recent decision of Hurlbut v. Conoco, Inc., 253 Kan. 515, 537, 856 P.2d 1313 (1993), "[T]he question of admissibility of scientific or experimental tests is one of relevance; viz., does the experiment have any tendency to prove or disprove a consequential fact?" See 22 Wright Graham, Federal Practice and Procedure, Evidence § 5171 (1978). The State argues that the experiment was relevant to impeach the defendant's theory. The defendant objects based upon his contention that the conditions are not the same or similar, making the evidence irrelevant.

In Hurlbut, we said that "[t]he admission of evidence of experiments must be established by showing background proof that the experiments were conducted under conditions that were at least similar to those which existed at the time of the incident in question." 253 Kan. 515, Syl. ¶ 10. We cited an earlier case of Spraker v. Lankin, 218 Kan. 609, 545 P.2d 352 (1976), where there was a challenge to the admission of accident reconstruction evidence. We said in Spraker that the following differences went to the weight of the testimony rather than its competency:

"Some of the obvious differences [between the test vehicles and test conditions and collision circumstances] are the grade at the accident scene and the level test site; the different makes, sizes, weights, types and construction of the motorcycles; the types of brakes; the types of tires; and the surface areas of the portions of the various vehicles which came in contact with the surface of the highway." 218 Kan. at 614.

As noted in Hurlbut, "The Spraker court concluded that such differences go to the weight of the testimony rather than its competency [citation omitted]. The court stated the differences should be recognized and may be explained by the expert. 218 Kan. at 614." 253 Kan. at 537.

In Timsah v. General Motors Corp., 225 Kan. 305, 591 P.2d 154 (1979), the plaintiffs contended that the court erred in refusing to permit them to conduct an in-court experiment to establish whether the hydraulic reservoir, a component of the steering system, would retain fluid when turned upside down. The defendant had objected to the experiment because there was no offer of proof that the conditions involved in the proposed experiment would be similar to the conditions which were present at the time of the accident. We said: "The relevance of demonstrations or tests to be performed in the presence of the jury rests in the sound discretion of the trial court and its decision on the matter will not be reversed on appeal unless an abuse of discretion is apparent." 225 Kan. at 317.

The defendant argues in this case that the following differences between the test conditions and the homicide conditions justified exclusion of the test evidence absent some reasonable explanation of the effects of the differences: There were windows in Mrs. Brown's house that were closed during the test that were open when the body was found; the volume of the neighbor's television could not rationally be evaluated; it was not clear that precisely the same type of ammunition was used during the test as during the homicide; because the murder weapon was not discovered, the same weapon was not used in the test as in the homicide; it could not be determined that the gun barrel was pushed into the pillow to the same depth during the test and the homicide; the outside temperatures were colder during the test than during the homicide; and the shots were fired into two telephone books placed on top of a wooden chair rather than a human head on a bed.

The State argues that the test conditions nevertheless were substantially similar to the conditions at the homicide. The test was conducted in the early morning hours; when Mrs. Brown's body was found, the south window in her bedroom was open to the same degree it was open during the test. Although the officer did not open the other windows that were open when the body was found, he testified that those windows faced west, and the neighbor's house was east of the victim's house. Accordingly, he did not believe the difference was significant. The test shots were fired through a pillow identical to that found at the scene in a location substantially the same as the body's location. A .22 caliber handgun was used in the test and the homicide. The neighbor was located in the same location in his home during the test as he was during the homicide.

The State also notes that other differences might increase the likelihood that the neighbor could hear the shots during the test. The air conditioner was turned off during the test firing; the neighbor was awake during the test, but awakened from sleep by the sounds he heard the night of the murder.

The law does not require that all conditions be replicated in a test such as this one. It appears that the State made every effort to find and use a weapon that police had reason to believe was substantially similar to the murder weapon. It was not possible to replicate outside temperature without waiting several months for summer to arrive. Although defendant claims the television volume could not rationally be evaluated, the neighbor candidly admitted the volume was "about the same" as it was on the night of the murder. Although shots were fired into telephone books instead of a human head, we do not believe that defendant expected the police to use a human head for the test.

The test evidence was relevant because it had a tendency to prove or disprove a consequential fact — whether the neighbor heard the fatal shots. The admissibility of this relevant evidence rests in the sound discretion of the trial court. Although we recognize that there were some differences in the conditions, the defendant has not demonstrated that the conditions were so dissimilar that admission would amount to an abuse of discretion. In fact, conditions during the test were substantially similar to the conditions on the night of the homicide.

(2) Jury Instruction: Must or Should?

The defendant next claims that the trial court erred in failing to instruct the jury that it must not draw inferences from the defendant's failure to testify. The court instructed the jury in accordance with PIK Crim.3d 52.13: "You should not consider the fact that the defendant did not testify in arriving at your verdict." The defendant requested that "should" be replaced by "must"; the trial court declined the requested modification.

In support of his contention that the trial court's failure to give the requested instruction was reversible error, the defendant cites Bruno v. United States, 308 U.S. 287, 84 L.Ed. 257, 60 S.Ct. 198 (1939), for the proposition that a defendant has a right to have the jury instructed that it must not consider a defendant's failure to testify. Bruno does not dictate such a result. In Bruno, the defendant had requested the following instruction:

"`The failure of any defendant to take the witness stand and testify in his own behalf, does not create any presumption against him; the jury is charged that it must not permit that fact to weigh in the slightest degree against any such defendant, nor should this fact enter into the discussions or deliberations of the jury in any manner.'" 308 U.S. at 292.

The Supreme Court concluded that the "topic on which Bruno proffered an instruction had not been charged at all." 308 U.S. at 292. Unlike in our case, no instruction on this subject was given in Bruno. Under these circumstances, the Court held that "the substance of the denied request should have been granted, and the judgment therefore is reversed." 308 U.S. at 294. Bruno does not require the mandatory "shall" language, but did require that the substance of the defendant's requested instruction should have been given.

This same question has been recently addressed by this court in State v. Owens, 248 Kan. 273, 807 P.2d 101 (1991). In Owens, we found "no prejudicial error with the use of `should not' under PIK Crim.3d 52.13 rather than `must not' in an instruction informing the jury not to consider the fact that defendant did not testify." 248 Kan. 273, Syl. ¶ 8. We adhere to our decision in Owens and conclude that the failure to give the instruction requested by the defendant did not constitute prejudicial error.

In Owens, the defendant did not raise a specific objection and we were applying the standard of whether the failure to give the instruction was clearly erroneous. Here, the defendant specifically requested the use of the word must instead of should. Our conclusion in Owens that the failure to give the instruction did not amount to prejudicial error stands. Yet, we are of the opinion that a better practice would be to give an instruction containing the word must instead of should. While both give sufficient direction, as we noted in Owens, 248 Kan. at 284, "should is not as strong as must", and we believe the stronger term is more appropriate and constitutes the better practice. We would also note that PIK Crim.3d 51.02 provides in part: "You should decide the case by applying these instructions to the facts as you find them." While use of "should" provides sufficient direction, we believe the stronger term "must" also would be a better practice under PIK Crim.3d 51.02.

We hold that PIK Crim.3d 52.13 provides sufficient direction for the jury in its consideration of the defendant's failure to testify, that it "properly and fairly state[s] the law as applied to the facts in the case," and that the jury could not reasonably have been misled by the instructions. State v. Norris, 244 Kan. 22, 23, 765 P.2d 1120 (1988). We further hold that the better practice would be for the trial judge in the context of this instruction to use the stronger term "must" instead of the term "should." However, we conclude that no prejudicial error occurred in the failure to give the instruction requested by the defendant.

(3) Aiding and Abetting Instruction

The defendant's third allegation of error is that the trial court violated his right to due process of law by instructing the jury on aiding and abetting. The defendant contends that the State did not charge him with aiding and abetting and presented insufficient evidence of aiding and abetting to justify giving the instruction. The defendant objected at trial to the aiding and abetting instruction.

We have held that the State need not charge aiding and abetting in the charging document in order to pursue an aiding and abetting theory at trial. See, e.g., State v. Smolin, 221 Kan. 149, 152, 557 P.2d 1241 (1976); State v. Motor, 220 Kan. 99, 102, 551 P.2d 783 (1976). If, from the totality of the evidence, a jury reasonably could conclude that the defendant aided and abetted another in the commission of the crime, then it is appropriate to instruct the jury on aiding and abetting. State v. Clemons, 251 Kan. 473, 486, 836 P.2d 1147 (1992).

The defendant argues also that it was the State's theory that Paul Brown paid the defendant to kill Mrs. Brown and that the defendant was the one who actually committed the killing. The same was true in Clemons, and we held that such conditions do not preclude the giving of an aiding and abetting instruction. In Clemons, 251 Kan. at 485-86, the State's theory was that Clemons pulled the trigger, and the prosecuting attorney even attempted during closing argument to refute the possibility that anyone else did the actual shooting. We held that the aiding and abetting instruction was appropriate even though the State's theory was that the defendant actually committed the killing because when the evidence was viewed as a whole, a jury reasonably could have concluded that Clemons aided and abetted another. The same finding applies here.

There was evidence from which the jury reasonably could have found that the defendant and Paul Brown were at the Brown residence at the same time. One witness testified that the defendant had not returned to the Motel 6 party by 3:00 or 3:30 a.m., and Paul Brown's van was seen in front of the victim's house around 3:00 a.m. The defendant and Paul Brown were seen together on the afternoon of the murder. Paul Brown and the defendant both had attempted to establish alibis for the night of the murder. Paul Brown owned a .22 caliber handgun, and the victim died of wounds from a .22 caliber weapon. Paul Brown first denied owning any gun and later admitted he had owned one, but told police it had been stolen. After the homicide, the defendant was seen with a weapon similar to the one Paul Brown owned, and the defendant threw it into the river soon after the murder. There was evidence from a which a jury reasonably could find that the defendant actually committed the murder or that he aided and abetted Paul Brown in the commission of the murder. The trial court did not err in instructing the jury on aiding and abetting.

(4) Cautionary Instruction

The defendant claims the trial court erred by failing to instruct on testimony of a paid informant. He did not request the instruction at trial. When a party complains on appeal that an instruction was not given, but the party did not request the instruction at trial, we will reverse only if the failure to give the instruction was clearly erroneous. State v. Perkins, 248 Kan. 760, Syl. ¶ 8, 811 P.2d 1142 (1991). "The failure to give an instruction is clearly erroneous only if the reviewing court reaches a firm conviction that if the trial error had not occurred there was a real possibility the jury would have returned a different verdict." Perkins, 248 Kan. 761, Syl. ¶ 8.

The testimony at issue is that of Brian Ramey, who shared a jail cell with the defendant. We note first that Ramey did not receive anything in exchange for his testimony but testified that he came forward out of disgust for what the defendant had told him. He therefore was not a paid informant. This fact undermines the defendant's claim of error. However, the defendant asks us to apply the rule requiring the paid informant instruction by way of analogy, apparently because the witness is testifying for the State. Although this fact is insufficient to raise the paid informant analogy, we think it appropriate to demonstrate that the defendant's contention is without merit.

The defendant told Ramey that he had shot the victim three times in the head with a .22 caliber gun, that he placed a pillow over her head, and that he was to be paid $500 for killing her. He told Ramey that he had been dropped off at the scene and was gone for about one-half hour. He told Ramey that he wiped the gun clean and threw it into the river.

The defendant argues that the case of State v. Fuller, 15 Kan. App. 2d 34, 802 P.2d 599 (1990), rev. denied 248 Kan. 997 (1991), should apply. In Fuller, the Court of Appeals held that it is an error for a trial court to refuse to give the "paid informant" instruction (PIK Crim.3d 52.18-A) when the defendant requests such an instruction, where the informant's testimony is wholly uncorroborated, and when the informant's testimony provides the sole basis for the defendant's conviction. 15 Kan. App. 2d 34, Syl. ¶ 4. Here, the defendant did not request a paid informant instruction, the informant's testimony was substantially corroborated, and the informant's testimony was not the sole basis for the defendant's conviction. Fuller does not apply, but even if we were to apply Fuller, the defendant would not benefit. Suffice it to say, the witness was not a paid informant, and the trial court did not err by refusing to give a paid informant instruction.

Affirmed.