Impact glasses from Belize represent tektites from the Pleistocene Pantasma impact crater in Nicaragua

Tektites are terrestrial impact-generated glasses that are ejected long distance (up to 11,000 km), share unique characteristics and have a poorly understood formation process. Only four tektite strewn-fields are known, and three of them are sourced from known impact craters. Here we show that the recently discovered Pantasma impact crater (14 km diameter) in Nicaragua is the source of an impact glass strewn-field documented in Belize 530 km away. Their cogenesis is documented by coincidental ages, at 804 ± 9 ka, as well as consistent elemental compositions and isotopic ratios. The Belize impact glass share many characteristics with known tektites but also present several peculiar features. We propose that these glasses represent a previously unrecognized tektite strewn-field. These discoveries shed new light on the tektite formation process, which may be more common than previously claimed, as most known Pleistocene >10 km diameter cratering events have generated tektites.

Average data with s.d. is reported in supplementary table 2 for three pooled belizites samples, five Pantasma rocks (P4, P6, P9, P11 and P15) and two glasses (P1 and P2; full data already published in 3 ).

Method 3: Sr and Nd isotopes
Neodymium isotopes were measured in dry mode using an Aridus 2 dessolvator with solution at 25 ppb. The samples were bracketed with the Rennes Nd standard and corrected for the accepted value of the standard (0.511963 4 ). The external reproducibility based on 30 standard measurements is better than 29 ppm. The Nd and Sm spiked aliquots were measured afterwards and spike deconvolution used the same formulation as in 5 . For Sr isotopes measurements, the samples were measured in wet mode, at 400 ppb. Samples were bracketed with the NBS987 Sr standard and samples were corrected according to the accepted value (0.710248 6 ). The external reproducibility based on 18 standard measurements is better than 27 ppm. Finally, Rb/Sr ratios were measured on the Agilent 7700 ICP-MS by using a calibration curve made of gravimetrically prepared Rb/Sr solutions.
Two samples from belizite were measured, BLZ3 and BLZ6, and from Pantasma one glass (P1), two rocks (P6 and P15) and one soil (P17). Details on these Pantasma samples can be found in 3 . In the supplementary table 1 and in Fig.6 is also indicated for comparison the data published in abstract form by 7 . This work has also reported Os isotopic ratio, compatible with mantle derived origin.

Method 4: Cr isotopes
The chemical separation was adapted from 8 and described in detail in 9 . It includes three separation steps of Cr on cationic exchange resin AG50W-X8 and was conducted at the Institut de Physique du Globe de Paris. Measurement used a filament exhaustion sample standard bracketing approach following the method described in 10 .
Chromium isotopic data are reported using the ε-unit, which represents the relative deviation in parts per 10,000 of 53 Cr/ 52 Cr and 54 Cr/ 52 Cr ratios from a terrestrial standard (NIST SRM 3112a Cr standard; see supl. Table B). In addition to breccia sample P5B, a USGS geostandard, BHVO-2, was analyzed during the same session to check for the accuracy of the method.

Method 5: Ar/Ar geochronology
For Curtin laboratory measurements the glass samples were irradiated for 3 hours alongside FCs standards 11 , for which an age of 28.294 Ma (± 0.13%) was used 12 . The discs were Cd-shielded (to minimize undesirable nuclear interference reactions) and In LSCE Argon laboratory one belizite sample (JC collection) and one Pantasma impact glass sample (PV1) were irradiated for 2 hours in the Cadmium-Lined In-Core Irradiation Tube (CLICIT) Oregon State University TRIGA reactor (USA). Interference corrections were based on the nucleogenic production ratios quoted in 12 . After irradiation, two ~50 mg aliquots of each sample were loaded into a copper sample holder and placed within a laser sample chamber fitted with Cleartan© viewport.
Detailed analytical procedures can be found in 13 . Each aliquot was incrementally heated using a 25 Watts Synrad CO 2 laser. Extracted gases for each step were purified for 10 minutes using two hot AP 10 and two GP 50 getters (ZrAl). Each argon isotope measurement consisted of 20 cycles of peak-hopping. The neutron fluence (J) value for each sample was calculated using co-irradiated Alder Creek sanidine standard (ACs : 1.1891 Ma, equivalent to an FCs standard age of 28.294 Ma 59 ) and a 40 K total decay constant of 14 . J-values are the followings: Belizite = 0.00052333 ± 0.00000418 (2σ), Pantasma = 0.00052417 ± 0.00000420 (2σ). The mass discrimination was monitored by analyses of air pipette aliquots throughout the sample analyses, relative to a 40 6 and supplementary Table 2).

Method 6: 10 Be preparation and measurement
A carefully weighed mass spiked with of a 3025 ppm ± 9 ppm in-house 9 Be carrier 16 was totally dissolved in 48% hydrofluoric acid and fumed in a PTFE beaker until dry. The precipitate was dissolved with nitric acid and the Be was purified by solvent extractions and alkaline precipitations. After being oxidized at 800 °C for 1 hour, the BeO was mixed with niobium powder. All measurements were standardized against the in-house STD11 standard 17 . Analytical uncertainties (reported as 1σ) include uncertainties associated with AMS counting statistics, chemical blank measurements, and AMS internal error (0.5%). Note 3: 10 Be results Instead of a small target soil contamination, one may propose two alternative origins for the 10 Be measured in belizites: 1) in situ production since fall in Belize. In situ production over 0.8 Ma is at most 1.8 Mat/g, assuming the sample remained exposed at the surface for this whole duration. Field observations suggest on the contrary that samples remained under a soil or surface formation cover of the order of one meter on average, corresponding to about one third of the maximum production.
2) Presence of 10 Be in the volcanic rock due to subduction processes. Indeed contamination by oceanic sediments produce magma rich in 10 Be. 21  supplementary Fig.2 Mass distribution of belizites compared to ivoirites a) and b): mass distribution (in g) of belizite>10 g in JC collection, compared to ivoirites in our database 22 . Log-log fit allows to estimate the fractal dimension. To compare with usually published size distribution, one has to multiply this number by 3. Ar/ Ar