Opinion by Judge NOONAN; Concurrence by Judge REINHARDT.
NOONAN, Circuit Judge:William L. Hodge appeals his conviction of wire fraud in violation of 18 U.S.C. § 1343 and of false statements in violation of 18 U.S.C. § 1001 by filing certain certifications to the National Science Foundation (NSF) as an NSF grantee.
FACTS
Hodge is an atomic physicist with over seventeen years experience in atomic physics, laser physics, plasma physics, and elec-tro-optical instrument design. He has designed soft x-ray instrumentation at John Hopkins University, has taken part in plasma spectroscopy experiments at Massachusetts Institute of Technology (MIT) and at the Princeton Plasma Physics Laboratory, and has been a consultant at Sandia National Laboratory on x-ray lasers. He is the author of over forty scientific articles in these fields. He was the owner,and president of High Energy Laser Associates (HELA) in Oakland, California, founded in 1987. On December 21,1989, HELA submitted a grant proposal to the NSF under NSF’s Small Business Innovation Research program. HELA proposed to develop “a neodymium-like soft x-ray laser.” The principal investigator was listed as Hodge himself. Spectroscopic experiments were to be performed at the Janus laser facility, Lawrence Livermore National' Laboratory, Livermore, California and at MIT., Peter Hagelstein of the Electrical Engineering and Computer Science Department at MIT was to be consultant and subcontractor; Hagelstein’s group at MIT was constructing a slab amplifier that would be used in the project. Michael Finkenthal of Johns Hopkins University and Hebrew University was to consult in the area of spectroscopic line identification. The grant request was for $287,275 for the project, which was to be completed in two years.
On August 31, 1990 the NSF awarded a grant of $250,994, which included an “MIT subcontract budget dated August 27, 1990,” to HELA for the proposed project. This grant, the relevant one here, was a sequel to an earlier proposal by HELA for Phase I of the same project, which had been funded by the NSF in December 1988 and successfully completed by Hodge in November 1989. The new award was effective September 1, 1990 and expired February 28, 1993. The first installment of the grant, $66,338, was applied for by Hodge on August 31, 1990. To obtain it, he executed a NSF form headed *1150“Request for Advance Or Reimbursement.” He cheeked both the Advance and Reimbursement boxes. Under the subheading, “Computation Of Amount Of Reimbursements/Advances Requested,” there was a column containing as its first item “Total program outlays to date.” No sum was entered here. Item “i” was “Federal share now requested.” $66,338 was entered. A separate heading read “Alternate Computation For Advances Only.” Nothing was filled in here. The next heading was “Certification.” Under it in print the following appeared: “I certify that to the best of my knowledge and belief the data above are correct and that all outlays were made in accordance with the grant conditions or. other agreement and that payment is due and has not been previously requested.” Opposite this statement was a place for the “certifying official” to sign. Hodge signed.
On November 27, 1990, Hodge signed an identical form, cheeking both Advance and Reimbursement boxes and requesting $39,-503. On February 21,1991, he did the same, checking only the Advance box and requesting $20,000. On April 5, 1991, he did the same, checking the Advance box and asking for $75,700. Again on September 12, 1991, on January 21, 1992, and , on March 12, 1992, he checked only “Advance”, requesting, respectively, $35,000, $14,000, and $453.
To carry out his project Hodge needed access to the laser at Lawrence Livermore. He had had such access for Phase I, and access had been tentatively approved for Phrase II. For reasons that can only be guessed from the record, the Lawrence Liv-ermore committee that scheduled access did not grant it to him although Hodge persistently sought access from August 1990 until February 1991. By mid-1991, again for reasons not in the record, Hodge’s badge, permitting him to be at the Janus facility, was revoked. He continued to discuss his project with Dr. Michael Finkenthal between September 1990 and February 1992. In August 1991, Hodge indicated to Finkenthal that he was moving his project to MIT; however, he did not do so. According to Finkenthal, there was “something wrong” with Hodge. Hodge was under stress related to his marriage. At one point Finkenthal found himself being driven by Hodge in the opposite direction from the conference they were trying to attend. Hodge made no report to the NSF on what he had done.
PROCEEDINGS
On November 16, 1995, Hodge was indicted on six charges of wire fraud in violation of 18 U.S.C. § 1343 and six charges of false statements to a government agency in violation of 18 U.S.C. § 1001. The wire fraud charges alleged that Hodge had devised a scheme to defraud the NSF and carried it out by false statements to the NSF that the grant funds provided to HELA “were being used solely in accordance with the grant conditions”; the wire transfers were those made by the NSF in payment of the grant beginning with the payment of December 3, 1990. The false statement charges were that, beginning with the certificate of November 27, 1990, Hodge had.falsely certified that “all outlays were made in accordance with the grant conditions or other agreement.”
The jury acquitted Hodge on the counts charging false statements made on November 27, 1990, February 21, 1991 and April 5, 1991 and on the related wire fraud counts. The jury convicted him of wire fraud and false statements on the basis of the certificates of September 12, 1991, January 21, 1992, and March 12,1992.
Hodge appeals.
ANALYSIS
Each certification which is the center of this case cannot be read in isolation from the single page document, “Request For Advance Or Reimbursement,” of which the certification is a subordinate part. As the title of the document indicates, the form is a request for either an advance or a reimbursement. On each of the documents resulting in a conviction Hodge checked the box titled “Advance.” The term “outlays”, of which the certification speaks, refers to “Total program outlays to date,” item “a” under “Computation Of Amount Of Reimbursements/Advances Requested.” On none of the forms *1151signed by Hodge was any amount entered under “Total program outlays to date.” Hodge made no representation to the NSF as to what these outlays had been. Hodge did not ask reimbursement for any outlays made. In the documents resulting in convictions, Hodge sought only advances.
That Hodge sadly failed to carry out the research he proposed is evident. That Hodge had a moral obligation not to take the grant money and a moral obligation to return what he took does not need demonstration. That the government could have sued Hodge civilly for unjust enrichment is undisputable. That Hodge presented a ease where the government could show how tough the government can be with a nonperforming grantee is clear. What is not evident is Hodge’s crime or crimes.
The first certificate executed by Hodge, the day before the grant became effective, was not charged by the government as a false statement or as giving rise to wire fraud. This request was for $66,338 as reimbursement and advance. All outlays were certified as made in accordance with the grant although no program outlays were recorded. The “Alternative Computation For Advances Only” was not filled out. It is apparent from this first certification that the form required by the NSF was simply a convenient way for a grantee to pull down portions of the grant oh schedule. Perfunctory compliance with the form’s specifications was the rule. Nothing was significant except the amount asked for. Only this amount and the Advance/Reimbursement boxes were filled in by the grantee. The request form as written and as used was not intended as a report by the grantee to the government of what the grantee had actually done or expended.
All the subsequent requests were like the first. No outlays of funds were reported. No certification was made as to what had actually been spent on the project. The forms used in the counts of conviction were clearly labelled as requests for advances. No representation was made that reimbursement was sought for any outlay made.
Grantees dealing with the government must turn square corners. But the government must turn square corners when it employs the heavy engine of the criminal law. The government here has failed to prove that any reasonable person could find that the three certificates Hodge was convicted of falsely signing made any false statement to the NSF.
REVERSED.