ALBERT CECIL MOORE, APPELLANT,

v.

STATE OF FLORIDA, APPELLEE

ALTENBERND, Judge.

Albert Cecil Moore appeals his conviction and sentence for capital sexual battery. The alleged victim is his granddaughter. We reverse the conviction and remand for a new trial because the state introduced evidence of sexual misconduct by Mr. Moore that involved his daughters and occurred over thirty years earlier. Although the four incidents described in this evidence share a significant similarity with the charged offense in that all incidents occurred in a familial context, there are insufficient additional similarities to satisfy the tests for admissibility established in Heuring v. State, 513 So. 2d 122 (Fla.1987), Feller v. State, 637 So. 2d 911 (Fla.1994), and Saffor v. State, 20 Fla. L. Weekly S335 (Fla. July 13, 1995).

The information alleges that Mr. Moore committed a single act of capital sexual battery in 1989 by digitally penetrating his granddaughter’s vagina. The granddaughter, who was 14 at the time of trial, testified that Mr. Moore fondled her breasts and digitally penetrated her vagina on at least one occasion in 1989. This conduct occurred at his home.

In addition to describing the incident charged in the information, the granddaughter also testified that similar incidents had occurred on approximately eleven other occasions. These incidents occurred when she was between the ages of 7 and 11. They occurred when she was alone with her grandfather at his home. They also involved digi tal penetration. We conclude that the trial court properly permitted this evidence to be considered by the jury under the Heuring standard.

The state dedicated a significant portion of its case to presenting evidence of Mr. Moore’s sexual misconduct involving his own daughters when they were children. Mr. Moore is over 70 years of age. The earlier events occurred in the mid-1950s.

During pretrial discovery, all three of Mr. Moore’s daughters testified that, in their childhood, they were victims of sexual abuse by Mr. Moore. The trial court properly excluded the testimony of the third daughter because the alleged acts were not sufficiently similar to the charged crime. Of the two daughters who testified, the first stated that when she was 4 or 5 years old, her father once fondled her breasts and asked her to touch his penis. This incident, incident 1 on the chart below, occurred when they were walking in the woods. She touched his penis only momentarily.

The second daughter testified concerning three incidents, incidents 2 through 4 on the chart. Incident 2 occurred when she was 9 or 10 years old. While home alone with her father, he lay on top of her in bed. At the time of trial, she remembered that her vagina hurt, but did not remember whether she had her clothes on or what conduct caused her vagina to hurt. Incident 3 occurred when she was 10 or 11 years old. She went to the barn on the family farm to tell her father that she had just purchased her first bra. She remembered that he removed the bra and fondled her breasts. On a final occasion, incident 4, when she was 12, she went to the basement with her father. He kissed her “like an adult” and then masturbated in her presence.

The trial court admitted this evidence without the benefit of the supreme court’s recent decisions in Feller and Saffor. In Feller, the court explained that Heuring requires greater similarity between the offenses than many trial judges had realized:

Although sexual battery on an underage child is a reprehensible offense, it is not so unique in itself that it should be uniformly admissible under section 90.404(2), Florida Statutes (1989). Nor does Heuring stand for the proposition that any evidence of sexual abuse of a child is per se admissible.

637 So. 2d at 916 (footnote omitted). The supreme court reversed the judgment in Feller because the charged offense and the prior incident had only two things in common: both were the same type of offense and both involved young girls.

In Saffor, the supreme court recently rejected both a “stringent similarity” and a “liberal similarity” analysis of the Williams rule in a familiar context. The court held that the familiar context described in Heur-ing is a significant similarity, but some additional showing of similarity must be demonstrated under all the circumstances of the case before such prior incidents are admissible. The majority opinion in Saffor does not provide any compulsory factors for use in this analysis or require any particular weight to be given to various factors. Although it is clear that the trial court has discretion concerning the admission of such evidence, as we develop a case-by-case approach it seems equally clear that the trial court should make a careful and reasoned comparison of the similarities and dissimilarities between the charged offense and the prior incidents.

At least in the context of Mr. Moore’s case, the following chart permits one to compare and contrast three important aspects of the charged offense and the prior incidents: (1) the victims, (2) the alleged acts of abuse, and (3) the event. It also evaluates factors affecting the reliability of the testimony describing the prior incidents. While the chart contained in this decision has proven useful to our analysis and may prove useful in other eases, we are not mandating the use of this approach. Likewise, we have intentionally considered the victim factors first, the factors relating to the alleged act of abuse second, and the factors relating to similarity of the event third because that order of importance seems appropriate in this case. We have turned to the reliability factors only when the other factors are not decisive. Whether this particular analytical structure will be helpful in future case-by-ease decisions is not an issue that we can resolve today. SIMILARITY OF SIMILARITY OF THE ALLEGED ACT OF VICTIMS_ABUSE

Relation-Very Very Dis-Gender Age ship Similar Similar Dissimilar similar

Charged F 11 Grand-Offense Daughter

Incident 1 F 4-5 Daughter X

Incident 2 F 9-10 Daughter X

Incident 3 F 10-11 Daughter X

Incident 4 F 12 Daughter X

SIMILARITY OF THE RELIABILITY OF EVIDENCE

EVENT

Elapsed

General Specific Prior Prior Time since Clarity of

Time Location Location Conviction Charge Priors Testimony

Charged Evening Home Bedroom

Offense

Incident 1 Day Woods No No 40 yrs. Good

Incident 2 Evening? Home Bedroom No No 36 yrs. Poor

Incident 3 Day Barn No No 36 yrs. Good

Incident 4 Day Home Basement No No 34 yrs. Good

A comparison of the victims in this case indicates they all share the same gender. The young age of the victim in incident 1 is probably a significant difference. The victims in the other three incidents are similar in age to the victim in the charged offense. The distinction between a daughter and a granddaughter does not seem dispositive to this court.

Comparing the acts of abuse is more difficult. In this aspect, the chart is admittedly somewhat subjective. Nevertheless, this factor seems very important. In this case, the incidents are not the same type of offense. Neither daughter claims that Mr. Moore ever digitally penetrated her. Indeed, because the second daughter’s testimony regarding her father laying on her is so vague, neither daughter provides evidence of a sexual battery. Their testimony concerning the incidents in the 1950s describes conduct that would be categorized as lewd and lascivious acts. Athough both daughters and Mr. Moore’s granddaughter testified that he fondled their breasts, this conduct was not the basis for the offense charged in this case.

Thus, the state relied extensively on testimony pertaining to lewd acts to convict Mr. Moore of capital sexual battery. The analysis in Feller would strongly suggest that the dissimilarity between the offenses is enough to demonstrate that the incidents do not share the necessary unique characteristic or combination of characteristics. See also Hayes v. State, 20 Fla. L. Weekly S296 (Fla. June 22, 1995) (similar fact evidence not allowed to prove identity where crime charged was first-degree murder, and prior act was simple assault, which charge was dropped); cf. Saffor, 20 Fla. L. Weekly at S337 (An-stead, J., concurring). The majority opinion in Saffor, however, may not make this difference dispositive without an examination of all the remaining circumstances of the incidents.

Comparing the time and place of the events demonstrates substantial differences between incidents 1, 3, and 4 and the charged offense. On the other hand, substantial similarity exists between incident 2 and the charged offense.

The comparison of these three aspects clearly requires this court to hold that the trial court abused its discretion by introducing evidence of incidents 1, 3, and 4. The differences in the alleged acts, coupled with the differences in the time and place of the events, overcome the similarity between the victims.

Incident 2 is a closer question. The fact that this prior incident is not an act of sexual battery is very important, but not a disposi-tive dissimilarity. Because the factors we have considered in the first three aspects of this comparison do not clearly weigh either in favor or against admissibility, we conclude that the reliability of the testimony concerning this incident becomes a significant aspect in this decision. The victim does not recall the incident in detail. This is understandable because the incident occurred a long time ago when the witness was a young child. There was no criminal charge resulting from the incident, and thus, there is no recorded testimony or public record documents to verify the event. Therefore, while it is entirely possible for thirty-year-old incidents to be used as evidence in this type of criminal case, we conclude that the significant dissimilarities in the acts of abuse, and the risk that the limited testimony describing this incident may not be reliable, require that this incident be excluded from evidence.

We have reviewed the record and unfortunately cannot conclude beyond a reasonable doubt that the erroneous admission of this evidence was harmless. See State v. DiGuilio, 491 So. 2d 1129 (Fla.1986). Therefore, we are required to reverse Mr. Moore’s conviction and sentence for capital sexual battery and remand the case for a new trial.

Reversed and remanded for new trial.

PARKER, A.C.J., and FULMER, J., concur.