Context. The Crab nebula was observed with the HESS stereoscopic Cherenkov-telescope array between October 2003 and January 2005 for a total of 22.9 h (after data quality selection). This period of time partly overlapped with the commissioning phase of the experiment; observations were made with three operational telescopes in late 2003 and with the complete 4 telescope array in January-February 2004 and October 2004-January 2005. Aims. Observations of the Crab nebula are discussed and used as an example to detail the flux and spectral analysis procedures of HESS. The results are used to evaluate the systematic uncertainties in HESS flux measurements. Methods. The Crab nebula data are analysed using standard HESS analysis procedures, which are described in detail. The flux and spectrum of γ-rays from the source are calculated on run-by-run and monthly time-scales, and a correction is applied for long-term variations in the detector sensitivity. Comparisons of the measured flux and spectrum over the observation period, along with the results from a number of different analysis procedures are used to estimate systematic uncertainties in the measurements. Results. The data, taken at a range of zenith angles between 45• and 65• , show a clear signal with over 7500 excess events. The energy spectrum is found to follow a power law with an exponential cutoff, with photon index Γ = 2.39 ± 0.03 stat and cutoff energy E c = (14.3 ± 2.1 stat ) TeV between 440 GeV and 40 TeV. The observed integral flux above 1 TeV is (2.26 ± 0.08 stat ) × 10 −11 cm −2 s −1 . The estimated systematic error on the flux measurement is estimated to be 20%, while the estimated systematic error on the spectral slope is 0.1.

The diffuse extragalactic background light consists of the sum of the starlight emitted by galaxies through the history of the Universe, and it could also have an important contribution from the first stars, which may have formed before galaxy formation began. Direct measurements are difficult and1 not yet conclusive, owing to the large uncertainties caused by the bright foreground emission associated with zodiacal light 1 . An alternative approach 2-5 is to study the absorption features imprinted on the γ-ray spectra of distant extragalactic objects by interactions of those photons with the background light photons 6 . Here we report the discovery of γ-ray emission from the blazars 7 H 2356−309 and 1ES 1101−232, at redshifts z=0.165 and z=0.186, respectively. Their unexpectedly hard spectra provide an upper limit on the background light at optical/near-infrared wavelengths that appears to be very close to the lower limit given by the integrated light of resolved galaxies 8 . The background flux at these wavelengths accordingly seems to be strongly dominated by the direct starlight from galaxies, thus excluding a large contribution from other sources -in particular from the first stars formed 9 . This result also indicates that intergalactic space is more transparent to γ-rays than previously thought.The observations were carried out with the High Energy Stereoscopic System 10 (H.E.S.S. ), a system of four imaging atmospheric Cherenkov telescopes operating at energies E ≥ 0.1 TeV. These two blazars are at present the most distant sources for which spectra have been measured at these energies (Tab. 1).Intergalactic absorption is caused by the process of photon-photon collision and pair production. The original spectrum emitted by the source (which we call "intrinsic") is modified such that the observed flux E) , where the optical depth τ (E) depends on the Spectral Energy Distribution (SED) of the Extragalactic Background Light (EBL) (Fig. 1). Details are provided in the Supplementary Notes and Figures. For any reasonable range of fluxes at ultraviolet (UV) and optical/near-infrared wavelengths (O-NIR), τ (E) -and thus absorption -is larger at 1 TeV with respect to 0.2 TeV. This difference makes the observed spectrum steeper (that is, Γ obs > Γ int , for a power-law model dN/dE ∝ E −Γ ) The spectral change ∆Γ=Γ obs − Γ int scales linearly with the EBL normalization, and becomes more pronounced at larger redshifts. Thus more distant objects provide a more sensitive diagnostic tool.In general, if the intrinsic spectrum were sufficiently well known, τ (E) -and thus the EBL SEDcould be effectively measured by comparing intrinsic with observed spectra. Blazars, however, are characterized by a wide range of possible spectra, and the present understanding of their radiation processes is not yet complete enough to reliably predict their intrinsic γ-ray spectra. But for these two sources, with O-NIR fluxes at the level of the "direct" estimates, the intrinsic spectra needed to reproduce the H.E.S.S. data become extremely...