Reviewer #2: I have read with interest the paper "Electromagnetic calorimeters based on scintillating lead tungstate crystals for experiments at Jefferson Lab", which is appropriate for publication on Nucl. Inst. Meth. A. The paper is in general clearly and well written and, therefore, my comments are relatively minor. General Uniformize the choice of the past tense used in the paper, e.g. on L 164-165 in the same sentence: "is used to monitor" …. "was read-out". Would suggest, "is read-out", and the whole paper should be consistently checked. Cross-check the use of "beam photon" and "photon beam" throughout the paper. Figures Y-axis, when appropriate, replace "Events" ==> "Events/bin width" writing the actual bin width. All figures showing fits to experimental data: use the same standard of Fig. 8, showing the fit parameters. Fig. 12: please check the appearance of the used colours in B/W. Same for fig. 15. Caption fig. 14: add elasticity definition. Comments to the text L92 as the calibration section only comes after a few pages, maybe add: "as discussed in Section 3.5" or similar. L106 for consistency with L97, add ref. L191 "induced by the" (add the) L217 add missing "%": 2.63% +- 0.01%, similarly for C. N in MeV? Add. L244 Geant refs. missing Summary Define acronyms used in the Summary e.q. L478 Compton calorimeter (CCAL). References: add missing DOIs Reviewer #3: This paper reports about the construction and performance of an electromagnetic calorimeter made by a matrix of ~1000 crystals of PbWO4 read by PMTs. A different arrangement of the same crystals makes possible (or will make possible) to use the calorimeter in different experiments: PRIMEX, GLUEX and NPS at Jefferson Lab. The authors report about the measured performance of the CCAL during the PRIMEX experiment as well as some other on-beam tests focused to check and optimize the experimental set up to face the different conditions and requirements of the other assemblies (GLUEX and NPS). While the report is well written and easy to read, there is a little of confusion between what has been done for the CCAL and what is related to the other two future projects. The text will benefit by a rearrangement of the different paragraphs with a clear separation between the three projects clarifying what has been done and what will be done. This confusion is particularly evident in the Introduction where the three projects are mixed making the reading difficult. The authors should also describe in more details the calibration procedure adopted for the CCAL since the final performance strongly depends on it. After the revision, the paper should be considered for publication since the information and procedures herby described are relevant for anyone interested in PbWO4 em calorimeters. Beside these general considerations, below the list of detailed comments (L=Line) Detailed comments Replace everywhere the word 'module' by 'crystals': it is misleading since by module usually one considers a matrix of crystals Abstract: forward calorimeter -> Forward CALorimeter L3-4: add (some) references to CMS, ALICE, PANDA, HPS, CLAS and CLAS12 calorimeters L5: I'd also add the radiation hardness as one of the main characteristic of PbWO4 L14: add (NPS) after the name L22: EIC acronym precedes the name explanation in L24 L29: FCAL acronym is not explained (other than in the Abstract) The whole Introduction is a little bit confusing: I understand that crystals were used in different detectors and arrangements but, it is not completely clear which of these many arrangements the authors are reporting about neither if the different detectors were built or are expected to be assembled in future. L38: Beam photons -> Photons L45: lepton -> electron, otherwise can be confused with positron production L62: reconstruction of particles -> particles reconstruction L63: " ... crystal and positioned it about 64 6 m downstream" -> crystal. The CCAL will be positioned .... L64: add 'Theta' to angle L79: dLY/dT should reflect the inverse proportionality of LY on T L85: I'm surprise that no insulator surrounds the active volume of the calorimeter. How the T was kept uniform in the volume? how the PMT heat (30W) was decoupled from the cold mass? L160: it would be nice to have some details about the LED signal and how it compares to the actual PbWO scintillation light L180: specify in which time frame the 6% stability has been measured L185: specify the (energy) range of the photon beam used to calibrate the detector L192: it is not clear what Fig.7 shows: is it the amplitude of a SINGLE crystal (seed of the shower?) or the energy of the CLUSTER (entire shower)? how crystals by crystals were equalized? Hpw is it possible that the energy is fully contained and measured in one-crystal/cluster? 12 GeV photons are not fully absorbed by a 20cm long PbWO calorimeter and the linearity reported in Fig. 7 does not seem to be realistic. L197: not clear how the energy resolution is obtained. If it refers to the energy absorbed by a single crystal exposed to the 6 GeB photon beam, it should be dominated by transverse and longitudinal leakage. Was it accounted and corrected for? Otherwise, if multiple crystals are involved and the transverse component of the EM shower is summed up to the seed energy, how the later crystals have been calibrated? the calibration procedure should be described in more details. L210: not clear how a 5% non uniformity in individual crystal resolution (not clear how it is defined, see previous comment) reflects in the overall energy resolution of the CCAL (the sums up many crystals in the EM shower) L217: the resolution strongly depends on the energy thresholds (seed and sides crystals): please report here the value. Since the threshold (per crystal, I guess) mentioned in different part of the article is different (L185 15 MeV, L249 30MeV) it would be good report how the resolution changes by varying the threshold. No mention of the effect of the systematic error induced by the energy calibration on energy threshold is given. A note on it would be desirable. L226: what is the result of this alternative calibration? The authors should report the final parameter obtained with the Compton calibration if different from what quoted in L217 L244: I'm curious to see how simulations reproduce trigger rate as a function of the threshold and how they compare to data. It would be good to add it to Fig. 11. L248: I guess 200kHz is the max rate per crystal (and not per module - see my initial comment about the confusion module/crystal). Iy is just a coincidence that this is the same value quoted in L249 but with a different threshold? L252: the signal sampled by the fADC seems to be ~50-60ns wide: at 2 MHz, i'd expect some pile up effect. The check on the PMT performance at that rate are not the only possible cause of a degraded resolution. It would be good to make a comment here. L269: did the author tried to fit the time dependence on the reconstructed shower energy? It would be nice to see the behavior. L272: is the deposited energy corrected by the longitudinal leak? if so, how? any comparison to simulations? L277: some more details about the accuracy of the algorithm should be provided together with the cluster multiplicity per event. L279: how Compton candidates have been selected? this seems an important part that should be described in details since the conclusion (L289) that the calorimeter is able to reconstruct Compton events is based on a clean Compton sample identified in some other way. L292: is 130 MeV resolution consistent and comparable to the resolution quoted in Eq.1? are there any MonteCarlo studies demonstrating that the observed resolution matched with the expected one? L349: this is certainly true but, on the other hand, the longitudinal component of B is more difficult to shield ... Fig. 16 Caption: you may want to say that markers represents different HCoil generated field. L351 'most sensitive ...' it seems that a worl such as 'part' is missed. The whole phrase should be rewritten for better clarity. L400: what about the attenuation induced by the 'silicon cookie'? L434: it is good to report estimates based on current test but the ultimate performance in term saturation should be assessed by an on-beam measurement in the final configuration. I'd suggest the author to rewrite the conclusion mentioning explicitly that his studies are only preliminary. L444: I would rephrase the statement considering the factorization proof is a little bit more complicated that what should be tested in a single experiment and certainly, I hope, 'the entire 12 GeV Jefferson Lab semi-inclusive deep-inelastic scattering program' does not rely on this single experiment.