*4OPINION OF THE COURT
LAWRENCE V. JOHNSTON, Circuit Judge.THIS CAUSE comes before the Court on an appeal from the State of Florida because of an adverse ruling on a Pre-Trial Motion to Suppress. The Motion to Suppress attacked the admissibility of an intoxilyzer printout because it had been adjusted to give its results to accuracy levels of hundredths when in fact its actual capability would have permitted results to be recorded in thousandths.
This deliberate adjustment by the operator is claimed to have deprived the Defendant of possible exculpatory evidence. The adjustment was precipitated by a preception that the Department of Health and Rehabilitative Services (HRS) rules require some minimum variation between successive tests.1 The alleged minimum variation between tests if + or — 0.02%.2 The defense attorney claims that this minimum variation could be violated when the machine is adjusted to read in hundredths when a thousandths printout might force a new test. If a variation of, say .025% was noted, the operator would give another test because, mathematically, .025% is greater than .02%.3
The point is well taken.
The operator in this case has undertaken to give less credibility to a machine than it deserves. Forcing the machine to print in hundredths *5when it could show greater accuracy creates a technological paradox. If the operator of the machine would allow it to do all that it is designed to do the worst penalty anyone would have to endure is having to run a third or possibly fourth test until consistency appears and meets any reasonable standard, even that of Form 1615.4 There is no allegation that a third or fourth test is expensive or cumbersome or impossible. Adjusting the machine or intentionally forgetting the third digit to disregard perceptible variations gives the appearance of fulfilling a predilection to perform two and only two tests.5 The trial and appellate judges are then, as here, asked to cure legally what could have been cured factually with little effort at the time of arrest. Moreover, to create the fiction that the machine operates with less accuracy than it does, permits the confusion over the effect of Form 1615 to be perpetuated.6.
Therefore, the burden of either clarifying the rules or forms or facts must rest with the State and any attempt to reduce accuracy of a machine to be used as evidence against the Defendant in a criminal proceedings must be viewed in the strictest light.7
Affirmed.
JUDGES JACKSON, EDWARD M. and MOXLEY, JOHN DEAN, JR. concur.
The enumerated “rules” 10-D 42.21, etc., do not anywhere establish such a requirement. The only reference to the variation appears at the bottom of Form 1516, Oct. 84. This is an awkward place to add a substantive requirement. We are required to read the Form, together with F.S. 120.52(14) in order to discover that any Form of the department “. . . which imposes any requirement. . .” shall be a “rule” also.
The Form reads, in relevant part “Note: The two (2) breath tests shall be run no more than five (5) minutes apart and the two results are acceptable if with -t- or — 0.02% If more than 0.02% difference a third (3) test or a blood (sic) test should be given and the results recorded.
The Assistant State Attorney argues in her brief that even a 0.02% variation is acceptable because of F.S. 316.1932(l)(b)l. This statute attempts to remedy “ . . . insubstantial differences between .approved techniques and actual testing procedures in any case ...” (emphasis added). Thus, the statute grants an officer leeway to make minor omissions or errors without affecting the admissibility of evidence. The statute should not be read to inflate an administrative standard in every particular case where a certain variation occurs. If H.R.S. or the State want to establish an outside limit, they should stand by it. There must be a cutoff somewhere. It would be easy enough to self impose a + or — 0.028 or + or — 0.020 standard, but no purpose is served by setting such a precise standard (+ or — 0.020) if, when it is exceeded, the agency is permitted to argue the excess as insubstantial. The agency should be expected to possess more foresight in setting its standards than the State Attorney gives it credit.
Trial attorneys could easily predict a defense argument to the jury of “What are they hiding?” whenever partial results are offered on any matter.
Proscrustes did the same thing. This mythological Greek giant disregarded the perceptible variations in people for his “one size fits all” iron bed. He chopped off the legs of tall people and stretched the short ones. They all fit but few were comfortable with the result.
The confusion apparently stems in part from Form 1615 itself and H.R.S. Form 1514, September 82, which indicate the machine warm up and calibration procedures for the thousandths mode.
See U.S. v Seijo, 514 F.2d 1357 (2nd Cir. 1975). Undisclosed evidence is material if its development by skilled counsel could have induced a reasonable doubt. Where the evidence could have “probative value” its suppression is harmful. Shuler v State, 163 So.2d 3 (Fla. 1964), cert. denied, 379 U.S. 892 (1964). Failure to disclose detection of order of alcohol on defendants’ breath was reversible error. U.S. v Dye, 221 F.2d 763 (3rd Cir. 1955).