norml10 - Page 8

 RUG POLICY FOUNDATION TEL:202-557-$O07 _Fi ° 25,9t 12:56 No.O05 P.O_
 8
 of an accepted medical use, the cmwerse is not true--that
 absent FDA approval, commonly accepted medical use
 cannot be proven, Id at 890, Nor carl we conceive of a
 reason the Administrator should be barred from employ-
 ing notions developed by a sister agency insofar as those
 notions serve {.he missions of both age. ncies° a
 Which brings us to the most troubling part. off the
 Administrator's decision--the part which we think obliges
 us to order a remand. Petitioners, ahnost, in passing, point
 out that, three of the factors iI_ the Administrator's eight-
 factor test appear impossible to fulfi]_ and thus must be
 regarded as arbitrary and capricious. Impossible require-
 mencs hnposed by an agency are pert_,ree unreasonable:
 "Conditions imposed by [the] order are ... unreasonable
 by virtue of being impossible _o meet." D.C. Transit S3,s.,
 Inc. v. WasMngton Met_vpolitan Area Transit Comm'n,
 466 Fo2d 394, 402 (DoC. Cir.}, carl denied, 409 UoS. t086
 (1972). These three factors are:
 (4) General availability of the substance and intbrma-
 tion regarding the substance and its use;
 -(5) Recognition of its clinical t_se in generally
 accepted pharmacopeia, medical re%fences, journals
 or textbooks;
 (8) Recognition and use of the substance by a sub-
 stantial segment, of the medical practit,ioners in the
 Uni_.ed States,
 Pet.it%nero argue that one cannot logically show that a
 drug enjoys general "availability" or "use" by a substantial
 segment of medical practitioners if the drug remains in
 Schedule i, One of the very purposes in placing a drug in
 aPetitioners insist that the prior decision of this court in this
 case. National Org. [or the Reform of Marijuana Laws (NORMIJ
 u. Drug En/orcement Administration, 559 F.2d 735, 750 n.65 (D.C.
 Cir. 1977), also implies that the agency was :in error by relying
 on the FDA standards. Bat this court merely insisted on the
 distincz nature of the FDA mid the DEA proceedings without as
 much as hintir_g that the tests for compliance with the two stat-
 utes must be mutually exclusive. See NOBML, 559 P.2d at 750.