The plaintiffs, Dale Angelle and Arlene Laws, appeal the trial court’s affirmance of the Racing Commission’s judgment finding that the plaintiffs violated the racing rules which prohibit the giving of “milkshakes”, i.e. sodium bicarbonate 1, to their horses prior to racing.
In February of 2000, the Stewards of the Fair Grounds ruled that “Rare Money,” a thoroughbred horse trained by plaintiff-appellant, Dale Angelle, which won the fifth race held on February 7, 2000, had excessive levels of carbon dioxide After the race, Rare Money tested,
“... Positive for an overage of blood total dissolved carbon dioxide (C02)”. On February 20, 2000 a hearing was held. As a result the Stewards concluded that this condition is adverse to the best interests of racing and adverse to the best interest of the horse in that it alters its normal physiological state. In accordance with directive N. 03-2000 [Emphasis added.] Dale Angelle is hereby fined the sum of one thousand ($1,000.00) dollars.... [T]he purse is ordered redistributed ...
ALL FINES MUST BE PAID WITHIN FORTY EIGHT HOURS.
IN ACCORDANCE WITH LAC OR LRS
*11554:172 STEWARDS; AUTHORITY, POWERS, AND DUTIES; FINES AND SUSPENSIONS
35:1719 MASKING DRUGS 35:1721 MODERN THERAPEUTIC MEASURES
35:17 TRAINERS RESPONSIBLE FOR CONDITION OF HORSE
35:1 CLASSIFICATION OF FOREIGN |,SUBSTANCE BY CATEGORY 35:1797 PENALTY GUIDELINES
On February 10, 2000, the quarter-horse, “The Village Queen,” trained by plaintiff-appellant, Arlene Laws, won the eleventh race at Delta Downs. The Stewards there also found a violation of directive No. 03-2000 for the same reasons and assessed the same penalties as those assigned against the plaintiff-appellant, Angelle.
On April 27, 2000, a hearing was held before the Louisiana State Racing Commission on the appeals by the plaintiffs from the decision of the Stewards. The Louisiana State Racing Commission affirmed the decision of the Stewards. The “Notice of Adjudication” from the Racing Commission describes its decision simply:
Upheld steward’s ruling; fined $1,000; redistributed purse.
Both plaintiffs filed petitions for judicial review with the Civil District Court for the parish of Orleans, which proceedings were consolidated.
The district court after hearing oral argument and receiving into evidence the proceedings before the Racing Commission, affirmed the decisions of the Racing Commission.
The plaintiffs appealed the adverse district court judgment to this Court and filed a joint brief.
In their first assignment of error, the plaintiffs contend that the trial court erred by applying the wrong standard of review. Plaintiffs argue that the trial court failed to conduct an independent judicial review of the administrative proceeding, citing La. R.S. 49:964(G).
The district court in its “Judgment With Reasons” stated that:
| s[I]n the absence of manifest error, the Court is bound to accept the facts found by the Board and then can only determine if the Board properly applied those facts to the legal rules and laws involved. [Emphasis added.]
How this Court would have decided the case is totally irrelevant.
This is not a correct statement of the standard of review. La. R.S. 49:964(G) provides in pertinent part:
The court may affirm the decision of the agency or remand the case for further proceedings. The court may reverse or modify the decision if substantial rights of the appellant have been prejudiced because the administrative findings, inferences, conclusions or decisions are:
[[Image here]]
(3) Made upon unlawful procedure;
[[Image here]]
(5) Arbitrary or capricious or characterized by abuse of discretion clearly unwarranted exercise of discretion; or
(6) Not supported by a preponderance of the evidence as determined by the reviewing court. In the application of this rule, the court shall make its own determinations and conclusions of fact by a preponderance of evidence based upon its own evaluation of the record received in its entirety upon judicial review. In the application of the rule, where the agency has the opportunity to judge the credibility of the *1156witnesses by first-hand observation of demeanor on the witness stand and the reviewing court does not, due regard shall be given to the agency’s determination of credibility issues. [Emphasis added.]
As of July 12,1999, La. R.S. 49:964 was amended by deleting reference to the “manifestly erroneous” standard of review.2 Now under La. R.S. 49:964(G)(6), the reviewing court determines whether the administrative findings are supported by a preponderance of the evidence. In doing so the reviewing court makes its |4own findings based on what it determines to be a preponderance of the evidence. Doe v. La. State Board of Medical Examiners, 2000-1987 (La.App. 4 Cir. 5/30/01), 788 So.2d 1234. The trial court heard argument of counsel only. There was no live witness testimony. Credibility calls by the agency are entitled to “due regard” where, as in the instant case, the agency had the advantage of first-hand observation of the demeanor of witnesses and the reviewing court did not. La. R.S. 49:964(G)(6).
Moreover, a complete reading of the district court’s “Judgment With Reasons” reveals that it is based on a reliance upon the credibility of the testimony of the chemist for the Louisiana State Racing Commission, Dr. Steven Barker, although the trial court does not refer to him by name. This decision of the trial court to credit the testimony of Dr. Barker is consistent with the decision of the Racing Commission which obviously was also based on Dr. Barker’s testimony in preference to that of the plaintiffs’ expert, Dr. Kline. La. R.S. 49:964(G)(6) mandates that the district court give “due regard ... to the agency’s determination of credibility issues.” It is clear that the district court afforded “due regard” to Dr. Barker’s testimony and all other findings made by the trial court flow from the district court’s correct decision to credit that testimony consistent with the actions of the Racing Commission. Therefore, the district court’s erroneous reference to “manifest error” was harmless.
At the Racing Commission hearing both Angelle and Laws gave self-serving testimony that they had not administered “milkshakes” to their horses. Additionally, it was stipulated that Cynthia Menard would testify that she was with Ms. Laws almost all day on the day that the milkshake was supposedly | ¡¡administered and that at no time did she observe anything that would indicate to her that a milkshake had been administered to Ms. Laws’ horse on that day.
Plaintiffs called Dr. Kevin Kline, an associate professor of animal sciences at the University of Illinois as their expert. The racing commission called Dr. Steven Andrew Barker, a professor at the Louisiana State University School of Veterinary Medicine and the chemist for the Louisiana State Racing Commission.
Dr. Barker testified that the blood samples taken from Angelle’s horse had a C02 level of 39.3 mill moles per liter. Laws’ horse had a C02 level of 39.2 mill moles per liter. The threshold limits established by the Racing Commission were 37 mill moles for non-bleeders and 39 mill moles for bleeders. Both horses were bleeders. Dr. Barker stated that an increase in the level of carbon dioxide in a racehorse prior to a race could increase the horse’s ability to complete a long race. The sodium bicarbonate reduces the lactate in the muscles and lessens fatigue. Dr. Barker further stated that he has tested over 1500 *1157horses. Of the horses tested, the mean level of C02 was 29 mill moles. He also testified that quality control standards are conducted before and after the testing to insure that the calibration was correct. Dr. Barker acknowledged that the machine used in the testing was not designed solely for analyzing C02 levels. Appellants in their fourth assignment of error complain that the machine used to determine the C02 levels, an Olympus Auto-analyzer, was not designed primarily to test for specific C02 values, but for clinical evaluations, i.e., for general levels of blood gasses indicative of the existence of problems such a s heart disease, kidney disease, etc. However, Dr. Barker testified that it was accurate for such purposes. Moreover, appellants’ expert, Dr. Kline, admitted that the machine he used in Illinois was also used for | (¡clinical evaluations and was not designed specifically to test for C02 values. We find no merit in this argument of the appellants.
Plaintiffs’ expert, Dr. Kevin Kline, focused on the invalidity of the test results. Dr. Kline is a recognized expert in the analysis of sodium bicarbonate in horses. He testified that the test results for the plaintiffs’ horses should not be considered as he felt that the quality assurance test results showed that the machine was not calibrated correctly. Dr. Kline stated that the results are not valid because the runs from the quality control samples revealed a deviation in excess of the parameters suggested by the machine’s manufacturer. The machine manual provides that “[e]sti-mates of precision based on the NCCLS recommendations are consistent with typical performance. The within run precision is less than 3 percent and total precision less than 5 percent on the AU 600.” The two quality control samples used by the State were 13 mill moles and 28 mill moles. However, in the quality control runs made on the morning of February 8, 2000, the machine registered a level of 12.7 on the control sample known to have a level of 13 mill moles. At the same time the sample known to have a level of 28 registered 28.4 mill moles on the machine. The levels obtained in the afternoon, after testing was completed, were 13.6 mill moles and 26.3 mill moles using the same control samples.
Dr. Kline testified that the percentages of variability from morning to evening exceeded the level of imprecision that should be tolerated for the machine used. However, Dr. Barker explained that only the morning sample is used to determine the validity of the test results. The afternoon testing is not considered accurate because the test samples have been sitting out at room temperature all day, i.e., they could no longer be expected to provide true readings of 13 and 28 ^respectively. The afternoon run is done merely to determine that the machine is still working. Based on Dr. Barker’s testimony, the variance in the test results from the morning to the evening becomes irrelevant and the variance between the morning machine readings and the known value of the test samples is within acceptable limits. Dr. Barker further explained that Dr. Kline did not have the benefit of knowing about the failure to use fresh test samples in the afternoon, but that “otherwise his statements are relatively accurate about the variability.” In other words, both experts conceded that no machine was precisely accurate and that it is important to determine the range of error.
Dr. Kline suggested that it is not adequate to run a test sample at a high end of 28 when the violation standard is 39. Dr. Kline explained that the machine’s margin of error could increase at higher levels, a problem referred to as “linearity.” However, Dr. Barker testified that the machine *1158had been tested for the existence of linearity and none had been found to exist:
This machine has been looked at in the clinical quality lab for linearity. Entire standards have been run. The algorithm in the machine to calculate the curve has been validated for purposes of conducting good laboratory practice studies by the clinical laboratory practice studies by the clinical laboratory to meet FDA regulations. So we do have confidence in this machine. This machine like the machines used in Illinois and a number of other racing jurisdictions, these machines are all about the same and have the same kind of accuracy and precision.
.... We believe that we should have confidence in the results.
Based on the foregoing, we find no error in the reliance by the Stewards, the Racing Commission and the District Court on the results of the testing machine.
|sAlthough the Racing Commission produced the quality control results for February 8, 2000, applicable to Angele’s horse, the Racing Commission failed to produce the quality control results for February 11, 2000, applicable to Laws’ horse. When Laws’ attorney called for the results, the attorney for the Racing Commission explained that Dr. Barker did not have the results with him.
There is no presumption that the machine should be considered accurate on any given day. If that were the case there would be no reason to do quality control tests every day. The results of those tests are within the control of the defendant and/or Dr. Barker acting for the defendant. The Racing Commission produced the results for Angelle’s horse. The test results on Laws’ horse exceeded acceptable limits by less than the amount by which Angelle’s did, thereby comparatively increasing the probability that Laws’ test results could not be relied upon to prove that the C02 level in her horse exceeded the acceptable level when the machine margin of error is taken into account. Dr. Kline testified that in Illinois, which also has a limit of 39, the results in the case for both horses exceeded 39 by such small amounts that they would have been rounded down to 39 and would not have been considered to be in violation of the limit. However, Dr. Barker testified that the 39 threshold level was so high as to contain a built-in safeguard against a test error adverse to the plaintiffs. Therefore, Louisiana has not adopted the rounding procedure employed in Illinois.
In the absence of the quality control data for February 11, 2000, the Racing Commission failed to prove its case against Laws by a preponderance of the evidence. Additionally, the race run by Laws’ horse was a mere 220 yards, a distance so short that endurance would not be a factor. In other words, it is undisputed that Laws would have nothing to gain by administering a milkshake to her |9horse for such a short race. Moreover, unlike Angelle, Laws had a witness, Ms. Cynthia Menard, whose stipulated testimony corroborated Ms. Laws’ testimony that no milkshake had been administered to Laws’ horse on the day in question.
Thus the case against Ms. Laws is much weaker than the case against Angelle. In fact, so much so that we are able to conclude at this point after reviewing the record as a whole that the preponderance of the evidence does not support the case against Laws.
The plaintiffs were charged with violations of LAC Rule of Racing 35:1.17193, *1159which prohibits the use of a harmless substance which has the ability to mask a harmful one, and LAC Rule of Racing 35:1.17214, which prohibits the use of modern therapeutic drugs on the day of a race with certain limited exceptions. Although neither rule specifically refers to the use of sodium bicarbonate, Dr. Barker testified that the administering of large quantities of bicarbonate would have a drug-like effect, “so I think the rules actually already cover that.”
Later in the proceedings before the Racing Commission, Dr. Barker testified that:
[I]t is recognized that the administration of milk shakes can alter urine pH which can alter the excretion of certain drugs where they become undetectable. In some racing countries, milk shakes were initially of concern for their effect on fatigue, but the fact that they were altering urine pH and perhaps causing certain drugs not to be excreted and thus able to be detected, so milk shakes also have a capacity to serve as masking agents.
| mThus, based on Dr. Barker’s testimony, the record allowed the Racing Commission to determine by a preponderance of the evidence that the plaintiffs violated LAC Rule of Racing 35:1.1719 and/or LAC Rule of Racing 35:1.1721.
The plaintiffs also contend in their second and fifth assignments of error that the Commission sought to penalize them for the alleged violation of a rule which had not been promulgated at the time of the alleged offenses. On January 15, 2000, the Commission issued a directive concerning the use of sodium bicarbonate and the testing for the excessive amounts. Plaintiff Angelle’s horse ran on February 8, 2000 at the New Orleans Fairgrounds. Laws’ horse ran at Delta Downs on February 10, 2000. A rule on the use of sodium bicarbonate was not promulgated until September of 2000. La. R.S. 49:953 sets forth the procedure an agency must use in order to promulgate a rule. The statute also allows the use of emergency rules. See La. R.S. 49:953(B). La. R.S. 4:154(A) and La. R.S. 4:197 require the Racing Commission to comply with the rule-making provisions of the APA. The failure to comply with these provisions prevents the application of the proposed rule. Dorig-nac v. Louisiana State Racing Commission, 436 So.2d 667 (La.App. 4 Cir.1983); Star Enterprise v. State Through Dept. of Revenue and Taxation, 95-1980 (La.App. 1 Cir. 6/28/96), 676 So.2d 827.
We will concede for purposes of argument that the rule against milkshak-ing was not in effect at the time the Racing Commission ruled against the plaintiffs in February of 2000. However, based on Dr. Barker’s testimony, milkshaking was already prohibited by the more general rules against masking drags and race day drags. Plaintiffs were notified that they were appearing before the Racing Commission to respond to charges under these more general rules. The finding of the Racing Commission that the plaintiffs violated the directive against Inmilkshaking, regardless of its effectiveness, per force, includes by necessary implication a finding that the plaintiffs violated the more general prohibitions. Plaintiffs do not challenge *1160the validity of those more general provisions.
The record of the proceedings before the Racing Commission contains a pre-hearing “Memorandum” dated April 10, 2000, addressed to “All Commissioners” from Kim Raines Chatelain, Assistant Attorney General, explaining to the racing commissioners the nature of the charges against the plaintiff, Dale Angelle. It lists the following Rules: 35:1.1719, 1721, 1735, and 1797. Neither the directive nor the rule relating specifically to milkshaking are even mentioned. In the caption to two letters from Angelle’s attorney to the executive director of the Racing Commission prior to the hearing, Angelle’s attorney references the aforementioned rules, but makes no reference to the directive and the rule specifically directed against milkshaking. It is clear from the above described documents as well as the transcript of the proceedings before the Racing Commission what the nature of the charges against Mr. Angelle were and that they did not rely on the specific anti-milkshaking directive and rule.
Additionally, while the body of the decision of the Stewards contains a reference to the anti-milkshaking directive, No. 03-2000, it concludes by stating that the action of the Stewards was taken in accordance with certain rules among which were 35:1.1719,1721,1735, and 1797.
Dr. Barker testified that:
Lasix has been shown to have approximately two points increase in the total dissolved carbon dioxide of blood. That is the reason we have two levels. The thresholds were established through an international conference in '94 of International Analysts and Racing Veterinarians. After several studies had been published in the scientific literature, after several conferences had been held, one of |12which was chaired by chairman of our department at LSU and hosted by Louisiana State University School of Veterinary Medicine in 1992 in New Orleans and, based on that information, the international thresholds were recommended and accepted by different racing states. 37 for nonbleeders and [39]5 for bleeders.6 [Emphasis added.]
Dr. Barker went on to testify that only 50 out of a million horses would have a value as high as 35, and that only one per million would have a level as high as 36: “By the time we get to levels of 37, 38, 39, that becomes even less likely.” He explained that a level of 39 could not be attained by contamination:
Q. Would you say then, is it my understanding that any test that shows a level elevated above 37 for nonLasix source and over 39 for a Lasix source inclusively rules out any other alternative or hypothesis other than the giving of a milk shake as responsible for that reading?
A. Well, until someone comes forward with some other scientific data that supports a problem of using the reading, I have to go on the weight of scientific information that has already been produced and published in the scientific literature. So we are talking not anecdotal information or suspicion, or other things, we are talking about scientific information that supports the fact that *1161the 37 and 39 threshold excluding false positives as best anyone can, and that that information does not falsely lead to the accusation that someone has tried to violate the rules.
Dr. Barker was cross-examined on this question by Laws’ attorney:
Q. Isn’t it true that the 39 level, threshold that you call it and the 37 threshold were results of scientific tests?
A. Yes.
Thus, the test results on Angelle’s horse proved that there was a violation of the rules against masking drugs and race day drugs (LAC Rule of Racing 35:1:1719 11sand LAC Rule of Racing 35:1:1721) based on international testing standards independent of any other directives or rules directed specifically at milkshaking.
Finally, Dr. Barker noted that the plaintiffs got a further break by having their horses tested an hour and a half after the race instead of two hours. Testing two hours after the race would give a higher and more accurate reading, but it was felt that it was too much of an imposition to ask that the horses be kept so long after the race for testing.
Clearly, a preponderance of the evidence in the record as a whole supports the ruling against Angelle.
For the foregoing reasons, we affirm the judgment against Dale Angelle and reverse the judgment against Arlene M. Laws and render judgment in her favor.
JUDGMENT AGAINST DALE AN-GELLE AFFIRMED; JUDGMENT AGAINST ARLENE M. LAWS REVERSED.
LOVE, J., concurs in part and dissents in part.
. This has the effect of raising the level of blood total carbon dioxide to impermissibly high levels. This can have the effect of expediting the removal of lactic acid from the horse’s muscles, thereby improving its endurance for long races.
. Administrators of Tulane Educational Fund v. Johnson, 2000-0297, p. 3 (La.App. 4 Cir. 4/4/01), 784 So.2d 769, 771, fn. 3, writ denied 2001-1757 (La. 10/5/01), 798 So.2d 970.
. LAC Rule of Racing 35:1.1719 provides that "[t]he use of any drug or substance, regard*1159less of how harmless or innocuous it might be, which by its very nature might mask or screen the presence of a prohibited drug as provided in the Rules of Racing is prohibited.”
. LAC Rule of Racing 35:1.1721 states that "Hull use of modern therapeutic measures for the improvement and protection of the health of a horse is authorized, however, no such medication will be used on the day of the race except as may be provided in Chapter 15.”
. The transcript says "39.9” at this point, but it clear in the context of Dr. Barker's testimony as a whole that the transcript should have read'"39.” We assume that the "39.9” is just a typographical error.
. Both of the horses involved in this case are "bleeders” on Lasix.