This is a petition to review an order of the Drug Enforcement Administration (DEA) of the Department of Justice, fixing the individual manufacturing and procurement quotas for 1975 of the drug phenmetrazine, which in dosage form bears the trade name “Preludin”. The drug’s legitimate use is in connection with measures to deal with short term obesity. As a stimulant, it is also used illegally, without a prescription. It is a “controlled substance”, subject to the Controlled Substances Act, 21 U.S.C. § 801 et seq.
The companies seeking review of the 1975 quotas are the sole manufacturer of the basic ingredient, Western Fher Laboratories (Fher), and two intervenors, the sole manufacturer of Preludin in dosage form, Ciba-Geigy Corporation (Ciba), and the sole distributor to wholesalers, Boehringer Ingelheim Limited (Boeh-ringer).1 The production and distribution system involving this product has elements of uniqueness. Fher produces the basic ingredient in only the month of June in each year. The drug in dosage form involves a sustained release reaction and Ciba’s manufacturing process normally requires some six months for each batch of the main dosage form. All of Fher’s product — phenmetrazine—is sold to Ciba, and all of Ciba’s product— Preludin — is sold to Boehringer, which then sells it to some 490 wholesalers, who in turn deal with the potential market of the nation’s 53,332 retail pharmacies.
Under 21 U.S.C. § 826(a) and (c), the Attorney General is required to establish each year’s production and individual manufacturing quotas for each class of controlled substance. In this case, these quotas apply only to Fher. Fher’s quota is determined only after Ciba’s procurement quota is established, adding an allowance for an inventory for Fher. Ciba’s quota is determined in the light of current estimates of the medical needs of the consuming market. The objectives of such quotas are “to provide for the estimated medical, scientific, research, and industrial needs of the United States, for lawful export requirements, and for the establishment and maintenance of reserve stocks.” 21 U.S.C. § 826(a). The standards governing the fixing of these quotas are pre*328scribed in 21 U.S.C. § 826(c), which we quote in the margin.2
The quotas at issue in this case were established by the Compliance Investigations Division (CID) of DEA in the following manner. In working out Ciba’s procurement quota, CID started with the total sales figure for 1974, expressed in terms of kilograms (kg.) of anhydrous base (the controlled ingredient).3 This figure was 3316 kg. To this was added a 50 percent inventory allowance, 1658 kg., making a total of 4974 kg. From this was subtracted the ending inventories of goods in process and finished product of both Boehringer (307 kg.) and Ciba (850 kg.) — a deduction of 1157 kg. The resulting figure, 3817 kg., was further reduced by 616 kg. — an estimate of the difference by which national sales by Boehringer (3316 kg.) exceeded the national quantity of Preludin prescriptions dispensed (2700 kg.) in 1974. The latter figure was the result of the National Prescription Audit, a “stratified random sampling” of some 800 pharmacies recurrently conducted by a company known as IMS America Ltd.4 The result of this reduction was a procurement quota of 3201 kg. CID communicated this quota to Ciba on April 22, 1975, noting its reliance on the National Prescription Audit estimate, its “continued questioning” of the unexplained difference and stating “In the event that a reasonable explanation for this difference can be provided, D.E.A. will review its decision.”5
Petitioners then sought and had extensive hearings before an administrative law judge, claiming that the new quotas lacked a rational basis and were arbitrary. Their principal contentions, there as here, were that a 50 percent, or six month inventory allowance, is too little to cover the requirements of both Ciba, with its six month processing cycle, and Boehringer, which should maintain a three month inventory; and that the deduction of the 616 kg. of Preludin not accounted for as prescription sales was improper since the figure itself is unreliable and, even if it is accurate, it represents essential inventory at the retail level.
Hearings took place in May. The administrative law judge issued his recommended rulings, findings and conclusions in early July. He rejected the 1975 sales forecast made by Boehringer, noting past errors in prediction and the fact that first quarter sales in 1975 were “slightly lower” than in 1974.6 He accepted the HEW judgment of a constant *329sales level as a sounder starting point. He then took note of the profitability of illicit trade in Preludin, the fact that excessive inventories at the retail level facilitate diversion, and the desirability, in regulating a controlled substance, of keeping the pipeline “lean”. He noted CID’s reliance on the National Prescription Audit “to keep an eye on the size of the inventories at the retail level”. He found CID’s method and formula “more likely to meet the objectives set out in the statute than the alternative proposed by petitioners”, holding specifically that application [of CID’s formula] to reliable data will accomplish the overall statutory purpose”. He did, however, find that certain data from IMS America purporting to show a large increase in purchases by drug stores were unreliable.7 The administrative law judge therefore recommended that CID redetermine the 1975 quotas without giving any consideration to the suspect data.
DEA’s Acting Administrator accepted the judge’s report and recommendations and ordered the redetermination on July 31, 1975. This, technically, is the order being reviewed. On August 20, CID notified petitioners that it had made the redetermination without considering the suspect data and that it adhered to its decision of April 22.
Petitioners identify three issues for our determination: (1) whether DEA’s order contained sufficient findings and reasons to support its conclusion that a 50 percent inventory allowance was sufficient for both Ciba and Boehringer; (2) whether the quotas set by DEA were supported by substantial evidence; and (3) whether there is substantial evidence supporting the deduction of the 616 kg. difference between 1974 sales and the estimate of 1974 prescriptions.8 We consider these in inverse order.
The last two issues are closely intertwined. The government contends that a decision upholding the propriety of deducting the 616 kg. would automatically foreclose any possibility of reversing DEA’s refusal to allow an additional three months (25%) inventory to cover Boehringer’s needs. It would be, the argument runs, senseless to thin out the pipeline by 616 kg. only to refresh it by over 800 kg. While acknowledging the prominence of the issue relating to evidence of diversion, we need not accept this preemptive reasoning and shall address both issues.9
Petitioners’ attack on the use of the National Prescription Audit estimate is twofold. They challenge its reliability and assert that no reason was given “why this difference^] which [they contend] represents essential inventories and reserve stocks within the distribution chain, should be subtracted from the quotas.” We conclude that the Audit’s reliability as a basis for DEA’s judgment in this case has not been impeached. It is conceded that it has a reliability factor of ± 5.1 percent.10 Ciba has used the Audit in its own planning. Although the sampling does not include prescriptions dispensed by hospitals, clinics, govern*330ment installations, or physicians, testimony indicated that Preludin is “overwhelmingly” sold by retail pharmacies. The exclusion from the sampling of 982 retail stores in Alaska, Hawaii, and Puerto Rico amounts to 1.8 percent of the total; the result of this exclusion could not adversely affect the reliability of the result.
When we analyze the possible significance of the disparity between Boehringer’s sales and drug store prescriptions— amounting, in 1974, to twenty percent of sales — we have two alternatives: diversion or inventory build-up. Petitioners are in the position of the unfortunate Hobson. Although the evidence of reported thefts and losses involves a very small amount, 29 kg., or 673,000 dosage units, there was testimony that Preludin was “one of the most popular of the abuse drugs”, that it has been “the principal problem” for retailers in Philadelphia, as well as of special concern in Houston, Kansas City, and Washington, and that illicit sales bring prices of up to $15 per dosage unit, compared to a prescription sale price of 30 to 35 cents per unit. This evidence supports the conclusion that substantial amounts of the 616 kg. are being illicitly diverted somewhere along the pipeline.
Petitioners’ assertion that most of the disparity represents inventory build-up meets two difficulties. The first is that this is unlikely. A survey Boehringer conducted of its wholesalers, see infra, was relied on by Boehringer to show the dangerously low status of their inventories. The testimony was that a wholesaler’s normal inventory was a month’s sales. Wholesalers, it was further explained, had a greater need for inventory than retailers, who are serviced by deliveries several times a week and generally maintain “a fairly even flow of inventory.”11 Moreover, in 1973, the National Prescription Audit revealed a similar, smaller disparity of 373 kg. by which sales exceeded prescriptions — a disparity for which the DEA made no reduction in the 1974 quota. If this was not illicitly diverted, then it remains in inventory and, when added to the 616 kg., amounts to 989 kg. — or an inventory build-up, above and beyond that which existed in 1972, of over three and one half months of sales at the 1974 rate.12 Assuming, however, that petitioners are correct and that the 616 kg. (and the 372 kg. in 1973) went into inventory, the result would be that, whatever difficulties Ciba and Boehringer may have in ensuring a steady flow of product to wholesalers, there would be ample stocks on hand at the wholesale and retail level to serve medical needs. To the extent, therefore, that petitioners rely on this argument, their case is flawed by this internal inconsistency.
We therefore hold that there was substantial evidence to support DEA’s reduction of the quota by 616 kg. As we indicated above, see n. 9, there might be situations in which, because of the interest in serving the nation’s legitimate medical needs, reducing procurement quotas to make the pipeline more lean could be too crude an instrument. The legislative history of the controlled substances legislation and the concern of the Congress over the diversion problem 13 suggests that for such a situation to arise there would have to be, at *331the very least, a clear and definite showing that legitimate medical needs were not being satisfied. In our view it is clear from the record that this point has not been reached in this case.
Petitioners — particularly Ciba and Boehringer — have made an excellent showing that they are being and have been inconvenienced by not being allowed a larger quota. During 1973 and at least most of 1974, Ciba and Boehringer jointly had the benefit of large inventories; in 1973 they began the year with a thirteen month inventory (calculated in terms of 1973 sales)14 and in 1974 they began the year with a nine month inventory (in terms of 1974 sales).15 By the end of 1974, the joint inventory had been worked down to 1157 kg. — a four month inventory of which Ciba had approximately 3 months (850 kg.) and Boehringer a little over one month (307).16
Having used whatever cushion lay in their part of the pipeline, petitioners foresaw the plight they would find themselves in after the end of 1974. They succeeded in getting some increase in their 1974 quota, but sought to demonstrate by flow charts how, during most of 1975, Boehringer would be unable to fill current orders. Because of Fher’s June production schedule and Ciba’s six month processing cycle, Boehringer would be reduced to allocating such product as it received from Ciba to filling all or parts of orders backed up for several months.
In order to show the predicament that prevailed with respect to its wholesalers, Boehringer conducted through its field staff a rough survey of inventory status of the wholesalers. Of its 490 wholesalers it was able to obtain responses from 230. Of these, for whatever reason, Boehringer could only show that, during all of 1974 and the first four months of 1975, only 58 wholesalers had been out of stock for one month or more of Ciba’s 75 milligram product and only 19 wholesalers had been out of stock for one month or more of Ciba’s 25 milligram product. The fragmentary data could have been taken as indicating that at most one out of four wholesalers had been out of one of the dosage forms over the past 16 months. Twelve of these months were in 1974, when Ciba and Boehringer had started the year with an exceedingly large inventory. There is no indication that running out of stock showed an increase in the first four months of 1975.17 And there are absolutely no data concerning retailers. All that petitioners’ major witness on this issue could recall was that there had been some “anxious phone calls”.18 Although all of this might well make a case study in disorderly marketing, it falls entirely short of showing that retail pharmacies and their legitimate customers were not able to purchase the Preludin they needed.
*332We therefore hold that there was substantial evidence to support DEA’s limiting the inventory allowance to 50 percent of sales. Preludin being a controlled substance, DEA had the obligation, when it found substantial evidence of broad scale diversion, to attempt to achieve a more Spartan pipeline, even though this might cause inconvenience to manufacturer and distributor.
We also hold that CID, in reconsidering the quotas after the Acting Administrator had returned the matter to it, did not act improperly. In adhering to its original April 22 decision, without any reliance on the tainted data, we cannot say that it acted arbitrarily. Indeed, we do not see what else it should have done within the constraints of time. It was already August when the quotas were given CID for reconsideration. Although it was deprived of additional corroborative data, it had seen and heard nothing at the hearings which indicated that its original judgment was in error.
The only step it could theoretically have taken would have been to utilize ARCOS, a computerized system for collating reports of inventories and transactions from all wholesalers and pharmacies. But the testimony indicated that as yet this system is not functioning effectively. In any event, it is clear that, as of August, it could not be useful in providing data without considerable further delay.
It does seem to us, however, that there may be limits to sole reliance on leaner pipelines to confront the diversion problem. In order not to run the risk that legitimate consumers may be, for a substantial period, unable to purchase a drug, DEA would be well advised to attempt to ascertain more directly the state of inventories at the wholesale and retail levels.
Finally, we are not persuaded by petitioners’ claim that the Acting Administrator’s order was fatally defective under § 8(b) of the Administrative Procedure Act, 5 U.S.C. § 557(c), by reason of inadequate findings and reasons. In specifically approving CID’s formula and in impliedly19 approving reliance upon the National Prescription Audit figure to implement the Congressional policy of dealing with diversion by reducing inventories, the administrative law judge had clearly manifested his reasoning. Cf. Burlington Truck Lines v. United States, 371 U.S. 156, 167, 83 S.Ct. 239, 9 L.Ed.2d 207 (1962). It would, we think, be pointless to remand for a finding that “an inventory allowance of more than 50 percent for Ciba and Boehringer would only cancel out the reduction, otherwise justified, of 616 kg. and is not required to meet the medical needs of consumers.” The issues have been clearly seen and presented.
The order of the Acting Administrator is affirmed.
. We shall refer to all these corporations as “petitioners”.
. . In fixing [manufacturing] quotas, the Attorney General shall determine the manufacturer’s estimated disposal, inventory, and other requirements for the calendar year; and, in making this determination, the Attorney General shall consider the manufacturer’s current rate of disposal, the trend of the national disposal rate during the preceding calendar year, the manufacturer’s production cycle and inventory position, the economic availability of raw materials, yield and stability problems, emergencies such as strikes and fires, and other factors.”
Procurement quotas are specifically addressed in 21 C.F.R. § 1303.12.
. The sales figure is actually that of Boehringer’s sales to wholesalers, Ciba’s sales to Boeh-ringer being viewed as intra-system transfers. While petitioners originally sought the use of a higher sales figure as a starting point, the Department of Health, Education, and Welfare (HEW) reported, as the result of its studies pursuant to its obligation under 42 U.S.C. § 242(a), that “The use of stimulants [amphetamine, methamphetamine, methylphenidate, and phenmetrazine] in 1975 should be the same as in 1974, data indicate relative constant use for the past two years.” Petitioners do not now dispute the use of 1974 sales as a starting point.
.Its report of total kg. of Preludin dispensed in the past three years, adjusted to reflect the anhydrous base of the drug, was as follows:
% Change
1972 1973 1974 ’74 vs. ’73
2971.9 2465.7 2700.7 + 9.5%
. Ciba’s procurement quota of 3201 kg., increased by a 50% inventory allowance, and reduced by Fher’s 1974 year end inventory, became Fher’s manufacturing quota.
. First quarter sales in 1974 were 617 kg.; in 1975, 564.5, a drop of 8 percent.
. This information was received by CID after its quota letter of April 22, 1975. Even though it did not affect the initial decision, it did play a part in CID’s ongoing process of consideration and, as a CID witness acknowledged, was “utilized in our reassessment”. It was properly specifically excluded by the administrative law judge in any further reconsideration.
. Both sides agree that we must apply the “substantial evidence” test in reviewing the legality of the quota.
. We can conceive of a situation where, although large illicit diversions were taking place, they might be so concentrated in a few stores or locales that it would not serve the overall medical needs of the country to rely on the blunderbuss of thinning the whole pipeline. Orderly and timely filling of wholesalers’ orders might justify a larger total inventory for Ciba and Boehringer to respond to general needs, with more discrete action being required to address the diversion problem. While this situation may occur in the future, it is not revealed by this record. See infra,
.This means that there is a 95% probability that the actual amount of phenmetrazine dispensed by prescription in 1974, estimated at 2700 kg., will be within the range of 2698.62 and 2701.38 kg.
. While a witness for petitioners testified that retailers, who experienced a 9.5% increase in prescription sales in 1974 might double their inventories, the logic of a 100% increase in the light of the availability of quick deliveries escapes us, as it did the administrative law judge.
. 1974 sales — 3316 kg. 3316 divided by 12 = 276. 989 divided by 276 = 3.6.
. In hearings preceding enactment of the Controlled Substances Act, Congress had been informed by John Ingersoll, the Director of the Bureau of Narcotics and Dangerous Drugs (DEA’s predecessor), that as much as 50% of such drugs as amphetamines and barbiturates were diverted into illicit channels, and that between 75 and 90% of what appeared in the illicit market had been legitimately produced. Hearings before the House Subcommittee on Public Health and Safety, Committee on Interstate and Foreign Commerce, No. 91-46 (Part 2, 1970), H.R.11701 and H.R.13743, pp. 704-5.
The approach of the new legislation was “designed to prevent an over supply of such substances, which creates the potential for di*331version.” Statement of John Ingersoll. Hearings before the Committee on the Judiciary, U. S. Senate (91st Cong. 1969) p. 243.
. Government’s Exhibit 18 showed ending inventory, as of Dec. 31, 1972, at 3111.99 kg., and sales for 1973 of 2839. 2839 divided by 12 = 236. 3112 divided by 236 = 13.
. Indeed the beginning inventory, if couched in terms of the prior year’s sales would be almost eleven months. Inventory, Dec. 31, 1973. 2557 kg. 1974 sales: 3316. 2557 divided by 276 (monthly sales, 1974) = 9. 2557 divided by 236 (monthly sales, 1973) = 11.
Part of the reason for this amplitude of resources lay in the allowance of a supplemental quota in 1973, to allow for a strike anticipated in 1974.
. Boehringer, it may be noted, had a larger inventory as of Dec. 31, 1974 than it did a year earlier, when it had only 105 kg. on hand.
. Indeed, CID’s Chief testified that, as of March 1975, a print-out of back orders received by Boehringer from its wholesalers had totalled only 92 kg.
. We take note of the fact that 21 C.F.R. 1302.12(d) allows any holder of a quota to apply at any time for an adjustment, should there be “occurrences since the issuance of the procurement quota”. Had there been evidence of wide scale shortages at the wholesale or retail level, or even hard data indicating their imminency during the late spring or summer of 1975, such could have formed the basis for an application for an adjustment.
. We say “impliedly” because the administrative law judge, in specifically recommending that IMS America’s suspect data not be considered when CID reconsidered the quotas, must have intended the Audit to remain as a source for consideration.