The literature

This page exists so that nobody has to take the article on trust.  Every source it rests on is listed here with what it actually measured, who wrote it, and which way it cuts.

It is deliberately not a reading list of things that agree with this site.  The two papers that decide the argument, Letsolo 2010 and Tsipouri 2008, are not about chromosome 2 at all and have never been cited by either side of the creation debate.  The creationist papers are tagged by what their data show, and two of Tomkins's measurements are tagged as confirmed because they were reproduced for the article.

How to read an entry

Measurement or model.  A count of repeat units in a reference sequence is an observation.  A date from a substitution clock is a model, and it inherits every assumption in it.  Both are useful and they are not the same kind of thing.

Direction.  Assigned by what the source shows, not by what its author believes.

Verified.  Whether this project read the text directly (full), only its abstract (abstract), or only through another source that quotes it (via).  Nothing is quoted in the main article unless it is marked full or abstract.

Field definitions

Coverage and its limits, stated plainly.  This catalogue covers the fusion literature from Yunis and Prakash 1982 to Yang 2026, the three Tomkins papers and Jeanson's popular piece, and the two papers on what real fusion junctions contain.  Six items could not be obtained in full and are listed at the bottom;  two of them, Miga 2016 and Kasai 2000, carry results the article leans on and should be fetched.

The fusion site

The origin of man: a chromosomal pictorial legacy

Yunis, Jorge J. & Om Prakash.  Science 215(4539):1525–1530, 1982.  https://doi.org/10.1126/science.7063861

camp: mainstream topic: background method: measurement direction: background verified: abstract

Key claim: high-resolution banding of orangutan, gorilla, chimpanzee and human chromosomes;  18 of 23 human pairs virtually identical to the proposed common ancestor.

The banding study that first proposed the fusion.  Its method is G-banding, not sequence, and Bergman and Tomkins are right to say so.  The article presents it as the prediction and the sequence papers as the tests.  Closed at Science;  the abstract was taken from NCBI E-utilities.

Origin of human chromosome 2: an ancestral telomere-telomere fusion

IJdo, J. W., A. Baldini, D. C. Ward, S. T. Reeders & R. A. Wells.  PNAS 88(20):9051–9055, 1991.  https://doi.org/10.1073/pnas.88.20.9051

camp: mainstream topic: fusion-site method: measurement direction: undercuts verified: full

Key figures: two allelic cosmids mapping to 2q13 and no other interstitial site;  two degenerate arrays head to head;  flanks 80% identical over 269 bp when inverted, 95% and 90% identical to sequence beside two cloned telomeres.

The discovery paper.  Two things matter for the creationist debate.  The word "degenerate" is IJdo's own, printed twenty-two years before Tomkins built an argument on the site's degeneracy.  And IJdo gave a mechanism for the shortness on the same page:  a long inverted repeat is unstable, forms cruciforms and is trimmed.  The paper gives no repeat count;  the 158 and 127 that circulate come from Fan 2002.  Read from the scanned PMC PDF, pages anchored by running heads at both ends.

Genomic structure and evolution of the ancestral chromosome fusion site in 2q13–2q14.1

Fan, Yuxin, et al.  Genome Research 12(11):1651–1662, 2002.  https://doi.org/10.1101/gr.337602

camp: mainstream topic: fusion-site method: measurement direction: complicates verified: full

Key figures: two head-to-head arrays with nothing between them;  14% diverged from canonical repeats;  48% of 127 and 46% of 158 units perfect;  flanks match 9pter and 22qter to within 21 kb and 1.4 kb;  the two human alleles differ by 12% within the array;  Kasai's 38 cosmids in the same order on human 2 and chimpanzee 12 and 13.

The most-quoted source on the creationist side, and genuinely two-sided.  It quantifies the degeneracy and calls it a puzzle.  It then answers the puzzle on the same page:  a join at already-degenerate arrays (since confirmed by Yang 2026), or rapid decay shown by the 12% allelic divergence.  The creationist literature quotes the question and stops before the answer.  The 12% is real:  the T2T-quality HG002 paternal haplotype reproduces IJdo's 1991 array to 0.1% (research/allele-compare.ts).  But five human copies fall into two haplotype groups that also differ by 1 to 5% in the flanking non-repeat DNA, more than human and chimpanzee differ, so the 12% is not point mutation since the alleles split and cannot date the fusion for either side.

Gene content and function of the ancestral chromosome fusion site

Fan, Yuxin, Tera Newman, Elena Linardopoulou & Barbara J. Trask.  Genome Research 12(11):1663–1672, 2002.  https://doi.org/10.1101/gr.338402

camp: mainstream topic: ddx11l2 method: measurement direction: undercuts verified: full

Key figures: the CHLR1 (now DDX11) copy at the fusion site shares only exons 18 and 22 to 25 plus the 3′ UTR with the full gene;  1,286 bp of exons 19 to 21 deleted and replaced by 40 bp;  the same fragment at 15qter, 16qter, 19pter and Xqter;  the intact 26-exon gene on 12p11.

The paper that settles the DDX11L2 argument eleven years before it was made.  The copy at the fusion site is a truncated subtelomeric duplicate present at four chromosome ends.  Tomkins cites the companion paper for the sequence and does not engage this one.

Generation and annotation of the DNA sequences of human chromosomes 2 and 4

Hillier, LaDeana W., et al.  Nature 434:724–731, 2005.  https://doi.org/10.1038/nature03466

camp: mainstream topic: cryptic-centromere method: measurement direction: undercuts verified: full

Key figures: junction at hg16 chr2:114,455,823–114,455,838;  a 2.6 Mb region at 2q21 enriched for pericentromeric duplications shared with ten chromosomes;  three tracts totalling 31,198 bp of alpha satellite over 36,696 bp.

The finished chromosome.  Cited for the first sequenced measurement of the dead centromere.  Its "chimp chromosome 13" is 2B in current numbering, a trap throughout this literature.  Free to read at nature.com.

The evolution of African great ape subtelomeric heterochromatin and the fusion of human chromosome 2

Ventura, Mario, et al.  Genome Research 22(6):1036–1049, 2012.  https://doi.org/10.1101/gr.136556.111

camp: mainstream topic: dating method: measurement direction: undercuts verified: full

Key figures: the segment hyperexpanded in chimpanzee and gorilla chromosome caps derives from human chr2:113,840,263–114,303,469 (build 35), beside the fusion site;  median 798 copies in chimpanzee, 1,092 in gorilla, fewer than 14 in every human and the Neanderthal;  the spread occurred more than 4 Myr ago.

A chromosome-end sequence in two apes is a two-copy interstitial block in humans, at the fusion point.  Ventura's dating argument is independent of any molecular clock and puts the fusion before 4 Myr, which conflicts with Poszewiecka.  The authors say their model is one of many possible explanations, and the article quotes that.  Ventura also shows subtelomeric material moving inland without fusion on chimpanzee VII and XIII, which is why the 2q13 case has to rest on architecture rather than on the mere presence of an interstitial repeat.

Incomplete lineage sorting of segmental duplications defines the human chromosome 2 fusion site

Yang, Zikun, et al.  Cell Genomics 6(1):101079, 2026.  https://doi.org/10.1016/j.xgen.2025.101079

camp: mainstream topic: fusion-site method: measurement direction: undercuts verified: full

Key figures: 109 kb site in T2T-CHM13v2.0 made of three segmental duplications from ancestral chromosomes 9, 12 and 22;  interstitial telomere sequences at chr2:114,027,333–114,028,207, 875 bp;  dead centromere at chr2:132,644,386–132,685,996, 41,610 bp, homologous to chimpanzee 2B, interrupted by SVA and L1PA3;  96.7 to 97.5% of its unique k-mers in Denisovan and Neanderthal;  no structural variants across 94 human haplotypes;  fusion placed at more than 5 Myr.

The strongest single source and the one to handle most carefully.  Its highlight "formed by SDs and subtelomeric repeats rather than telomeric sequences" will be quoted as if it denied the fusion;  the paper is about how the fusion happened, and the article prints the highlight and the model together.  It concedes to Tomkins that the site is transcribed and that the repeats predate the fusion, and refutes him on DDX11L2, the centromere, and the archaic sharing.  CC BY 4.0.

What real fusions leave

Comparative sequence analyses reveal sites of ancestral chromosomal fusions in the Indian muntjac genome

Tsipouri, Vaia, et al.  Genome Biology 9(10):R155, 2008.  https://doi.org/10.1186/gb-2008-9-10-r155

camp: mainstream topic: real-fusions method: measurement direction: undercuts verified: full

Key figures: seven ancestral fusion sites, 1.88 Mb of finished sequence with no gaps across the repeats;  retained (TTAGGG)n blocks of 168 to 837 bp;  centromeric satellite I beside every telomeric block.

The only published set of independently sequenced ancestral mammalian fusion sites.  The largest retained block, 837 bp, is the size of the human array.  It also supplies the point on which Tomkins is right, that known fusion sites carry satellite;  muntjac fusions are telomere-to-centromere, so the satellite belongs to the consumed centromere, and the human telomere-to-telomere event has no reason to carry it.  Never cited by either side of the creation debate.

The nature of telomere fusion and a definition of the critical telomere length in human cells

Capper, Rebecca, et al. and Duncan M. Baird.  Genes & Development 21(19):2495–2508, 2007.  https://doi.org/10.1101/gad.439107

camp: mainstream topic: real-fusions method: measurement direction: undercuts verified: full

Key figures: 47% of sequenced fusion junctions with no TTAGGG at all;  the rest between 0.33 and 12.8 repeats, mean 4.7;  the threshold below which a human telomere becomes fusable is under 13 pure TTAGGG repeats distal to the variant-repeat zone;  fusion detected in normal proliferating and senescent fibroblasts at about 4 per million cells;  TRF2-knockdown fusions, by contrast, "usually contain many kilobases of telomeric repeat DNA".

The measurement that answers the creationist argument at its root, and neither side has cited it.  Tomkins expects a fusion between two long capped telomeres to leave thousands of intact repeats.  Capper shows a telomere with thousands of intact repeats is bound by TRF1 and TRF2 and does not fuse.  The only known way to fuse long telomeres is to knock out TRF2 in the laboratory, which Capper names in the introduction.  Obtained during verification with bin/pmcget.py from PMC1993879 and added to the source library.

Fusion of short telomeres in human cells is characterized by extensive deletion and microhomology

Letsolo, Bo T., Jan Rowson & Duncan M. Baird.  Nucleic Acids Research 38(6):1841–1852, 2010.  https://doi.org/10.1093/nar/gkp1183

camp: mainstream topic: real-fusions method: measurement direction: undercuts verified: full

Key figures: 298 sequenced telomere-telomere fusions in human cells;  every one deleted at least one telomere, mean 2.44 kb;  38 to 60% with no repeat at the junction;  mean 5.8 repeats otherwise, maximum 43;  one of 298 with repeats on both sides;  86% of fusions inside the degenerate variant-repeat zone;  fusions observed with the 2q14 interstitial array itself.

The control experiment the whole argument needs.  Against it, chromosome 2 has about 25 times too much repeat for a fresh fusion, and the one configuration seen once in 298.  The cells were in crisis and the article says so;  crisis is the most permissive condition and the bias runs against the creationist case.  Never cited by either side of the creation debate.  CC BY-NC 2.5.

Comparative FISH mapping of the ancestral fusion point of human chromosome 2

Kasai, Fumio, et al.  Chromosome Research 8(8):727–735, 2000.  https://doi.org/10.1023/a:1026745505264

camp: mainstream topic: gene-order method: measurement direction: undercuts verified: abstract

Key figures: 38 cosmids from human 2q12–q14 ordered by high-resolution FISH on human chromosome 2, chimpanzee 12 and 13, and cynomolgus monkey 9 and 15;  no difference in relative order between human and chimpanzee;  the order inverted in each cynomolgus homologue.

The physical gene-order test, done five years before the draft chimpanzee assembly existed and by a method that cannot be scaffolded onto a reference, which is exactly Jeanson's objection.  The cynomolgus inversion is the control Fan does not report:  the method detects a difference in order when there is one.  Full text closed at Springer;  Unpaywall and Semantic Scholar both record the work as closed, and the abstract states the result in full.

The cryptic centromere and the archaic genomes

Evidence for an ancestral alphoid domain on the long arm of human chromosome 2

Avarello, R., et al.  Human Genetics 89(2):247–249, 1992.  https://doi.org/10.1007/BF00217134

camp: mainstream topic: cryptic-centromere method: measurement direction: undercuts verified: abstract

Key claim: low-stringency alphoid probes light every centromere and one other place, 2q21.3–q22.1.

The publication of the result IJdo cited as unpublished data.  A prediction that came true.  Abstract from NCBI E-utilities;  Springer paywalls the body, and the abstract carries the whole claim.

Chromosome-specific centromere sequences provide an estimate of the ancestral chromosome 2 fusion event in hominin genomes

Miga, Karen H.  Journal of Heredity 108(1):45–52, 2017.  https://doi.org/10.1093/jhered/esw039

camp: mainstream topic: cryptic-centromere method: measurement direction: undercuts verified: abstract

Key claim: both the active and the degenerate chromosome-2 centromere profiles are present in Denisovan and Neanderthal genomes.

The cleanest clock-free constraint in the topic, and only the abstract could be read.  Ten routes were tried during verification and all failed:  curl on the publisher PDF (403 behind Cloudflare), bin/cdp-fetch.py on the article page (redirects to the abstract) and on the PDF Unpaywall lists as free (returns paywall HTML), Unpaywall, Semantic Scholar, OpenAlex, Europe PMC, eScholarship, the Wayback Machine and Crossref.  The paper is genuinely closed.  The abstract carries the archaic-centromere claim in the author's own words, and nothing beyond the abstract is cited.  Miga also concedes that the fusion site is enriched for sequence shared elsewhere in the genome, which is Tomkins's observation in a mainstream paper.

A high-coverage genome sequence from an archaic Denisovan individual

Meyer, Matthias, et al. and Svante Pääbo.  Science 338(6104):222–226, 2012.  https://doi.org/10.1126/science.1224344

camp: mainstream topic: archaic-genomes method: measurement direction: complicates verified: full

Key figures: 12 sequenced Denisovan DNA fragments contain the head-to-head telomere junction;  chimpanzee and bonobo reads yield none;  the authors note that a karyotype difference "would likely have reduced the fertility" of Denisovan and modern-human offspring.

The primary for the claim the article had been taking through Poszewiecka.  It is a clock-free floor on the fusion's age for the mainstream, and it costs a young-earth model nothing if Denisovans are post-Flood humans.  Twelve fragments is a small number and the paper claims no more;  it is not an assembly across the fusion site.  Already in the source library from the Neanderthals and Denisovans article.

Revised time estimation of the ancestral human chromosome 2 fusion

Poszewiecka, Barbara, Krzysztof Gogolewski, Paweł Stankiewicz & Anna Gambin.  BMC Genomics 23(Suppl 6):616, 2022.  https://doi.org/10.1186/s12864-022-08828-7

camp: mainstream topic: dating method: model direction: complicates verified: full

Key figures: 0.9 Myr, 95% interval 0.4 to 1.5, from the deficit of GC-biased substitution clustering at the former telomere;  Dreszer's 0.74 Myr (0 to 2.81);  a bonobo control giving 0.67 Myr where no fusion happened;  the Denisovan genome carries the fusion.

The substitution clock, and the paper that puts the mainstream in visible disagreement with itself.  Its upper bound cannot be reconciled with Ventura or Yang, and its bonobo control gives a date where none should exist.  The article reports all of that and notes that the floor of the whole range is 66 times a young earth.  CC BY 4.0.

The creationist case

The chromosome 2 fusion model of human evolution, part 1

Bergman, Jerry & Jeffrey Tomkins.  Journal of Creation 25(2):106–110, 2011.  https://creation.com/images/pdfs/tj/j25_2/j25_2_106-110.pdf

camp: creationist topic: yec-argument method: review direction: complicates verified: full

Key claim: most fusion evidence predates genome sequencing and rests on banding and hybridisation.

Considered and not cited in the article.  The claim was already dated when published, since Fan 2002 is sequence, and Yang 2026 works entirely from finished assemblies.  Its lasting value is the reminder that Yunis and Prakash is a banding study, which the article states.

The chromosome 2 fusion model of human evolution, part 2

Tomkins, Jeffrey & Jerry Bergman.  Journal of Creation 25(2):111–117, 2011.  https://creation.com/images/pdfs/tj/j25_2/j25_2_111-117.pdf

camp: creationist topic: yec-argument method: measurement direction: complicates verified: full

Key figures: 798 bp core sequence;  45,450 TTAGGG and 45,770 CCCTAA across chromosome 2 at 40.2% GC;  no BLAT hit on chimpanzee 2A in panTro 2.1;  2q21 alphoid clones in three groups, not clustering with functional centromere clones.

The technical core of the creationist case and the best-argued creationist treatment.  Its 798 bp core matches hg38 base for base, and its hexamer counts were reproduced to within half a percent for the article.  Its central premise, that the site is pericentromeric with suppressed recombination, is false:  17.6 Mb from the centromeric bands at 0.85 cM/Mb, 33 times a genuinely pericentromeric window (length-weighted deCODE map;  an earlier unweighted figure of 4.56 was wrong).  Its chimpanzee negative result was a result about a 2011 assembly.

Alleged human chromosome 2 'fusion site' encodes an active DNA binding domain

Tomkins, Jeffrey P.  Answers Research Journal 6:367–375, 2013.  https://answersresearchjournal.org/alleged-human-chromosome-2-fusion-site/

camp: creationist topic: ddx11l2 method: measurement direction: complicates verified: full

Key claims: the array is in the first intron of DDX11L2;  CTCF, c-Myc and Pol II bind over it;  expressed in 255 cell and tissue types;  no synteny with chimpanzee 2A or 2B in Costa's FISH.

The strongest creationist argument.  Its observations were confirmed for the article against hg38 RefSeq.  Its inference fails because DDX11L2 is a non-coding three-exon fragment of a subtelomeric family, shown by Fan 2002b and Costa 2009, both of which predate the paper.  Its statement that Costa mentioned nothing of the fusion sequence is contradicted by Costa's Figure 4 legend.  Its Figure 3 caption calls the locus a pseudogene while the abstract calls it a functional gene.

An update on chromosome 2 'fusion'

Jeanson, Nathaniel.  Acts & Facts 42(9):13, 2013.  https://www.icr.org/content/update-chromosome-2-fusion

camp: creationist topic: yec-argument method: commentary direction: undercuts verified: full

Key claims: the degeneracy makes the evidence weak;  chimpanzee gene order was never independently determined.

The popular ICR form of the argument.  The gene-order claim was answered by Kasai's FISH in 2000 and is now moot with the T2T ape assemblies.  Cited for the popular form of the argument, not for any datum.

Debunking the debunkers

Tomkins, Jeffrey P.  Answers Research Journal 10:45–54, 2017.  https://assets.answersingenesis.org/doc/articles/pdf-versions/arj/v10/human_chromosome_2_fusion.pdf

camp: creationist topic: cryptic-centromere method: measurement direction: undercuts verified: full

Key figures: DDX11L2 now "long noncoding RNA gene";  the array as a second promoter with CAGE data;  the dead centromere at 41,608 bases with an L1PA3 of 5,957 and an SVA-E of 2,571, inside ANKRD30BL.

Tomkins's reply to critics, and the version of his case the article engages.  It measured the dead centromere and got the same answer the T2T consortium got nine years later, to two base pairs, with the same two named insertions.  It does not engage Letsolo, Tsipouri, Ventura, Avarello or Miga's archaic result and keeps the recombination-suppression premise.  Tagged undercuts because its centromere measurement confirms the mainstream description.  Its p. 50 argument that the relic "does not closely match homologous chimpanzee centromeres" rests on 1995 and 1997 hybridisation work and a BLASTN search of one consensus monomer against panTro;  Yang 2026's finished chimpanzee 2B centromere shows the homology.  Cited for the homology argument and its gap complaint.

DDX11L2 and fixation

DDX11L: a novel transcript family emerging from human subtelomeric regions

Costa, Valerio, et al.  BMC Genomics 10:250, 2009.  https://doi.org/10.1186/1471-2164-10-250

camp: mainstream topic: ddx11l2 method: measurement direction: undercuts verified: full

Key figures: the family maps to 1pter, 2q13/14.1, 2qter, 3qter, 6pter, 9pter/9qter, 11pter, 12pter, 15qter, 16qter, 17pter, 19pter, 20pter/20qter, Xpter/Xqter and Yqter;  derived from a rearranged portion of DDX11 propagated by segmental duplication.

The paper that named DDX11L2 and that Tomkins cites for it.  Every address is a chromosome end except the fusion site, and the Figure 4 legend names the fusion.  Costa does not claim a demonstrated protein.  A quotation attributed to this paper in Eric's notes is not in the paper and is in quarantine.  CC BY 2.0.

Cytogenetics and reproduction of sheep with multiple centric fusions (Robertsonian translocations)

Bruère, A. N. & P. M. Ellis.  Journal of Reproduction and Fertility 57(2):363–375, 1979.  https://doi.org/10.1530/jrf.0.0570363

camp: mainstream topic: fixation method: measurement direction: supports verified: abstract

Key figures: 856 ewes mated across 48 chromosomally different groups over ten years;  fertility not significantly different from New Zealand national breeding data;  under 1% phenotypically abnormal lambs in 1,995 progeny;  a triple-homozygous ewe and ram at 2n = 48 against the normal 54.

The strongest published evidence that chromosome number varies inside a kind at no measurable cost, which is the creationist side's best point in this topic.  It says nothing about the age of the human fusion.  Full text closed:  rep.bioscientifica.com 403s curl and bin/cdp-fetch.py on the DOI lands on an Oxford University Press abstract behind a payment wall.  The abstract carries every sentence cited.

Combinatorial genomic data refute the human chromosome 2 evolutionary fusion

Tomkins, Jeffrey P.  In Proceedings of the Eighth International Conference on Creationism, ed. J. H. Whitmore, pp. 222–228.  Pittsburgh:  Creation Science Fellowship, 2018.  https://cedarville.tind.io/record/21144/files/22_Tomkins_Chromosome_2_final.pdf  Library key tomkins-2018d.

camp: creationist topic: yec-argument method: measurement direction: complicates verified: full

Key claims: natural mammalian fusions all involve satellite DNA and centromeric breaks (p. 222);  cancer-cell telomere fusions are products of failed genome maintenance (p. 222);  the 798 bp core should be 98.5% like a pristine fused array but is 70% (p. 222);  the relic is 33,080 bases of satellite against 250,000 to 5,000,000 for working centromeres (p. 224);  it extends across ANKRD30BL exons (p. 224);  telomere-like repeats genome-wide overlap 5,197 Gencode features, 5,002 enhancer-TSS regions and 4,489 ReMap binding sites (Table 1, p. 226), proposed as designed regulators.

Tomkins's latest technical paper on the fusion, missed by the first draft and now cited.  The 70% reproduces in kind (84 to 85% identity by EMBOSS scoring, research/allele-compare.ts);  the 98.5% expectation assumes a pristine join that Letsolo's fusions and Yang's pre-fusion repeats rule out.  The exon claim is not borne out by current hg38 GENCODE and RefSeq tracks:  the satellite (chr2:132,208,801–132,250,410) holds no ANKRD30BL exon and the coding transcript NM_001358416.1 ends 47 kb short.  It shows Tomkins did ask what natural fusions leave, though not how much repeat a telomere junction keeps.  The ENCODE intersection supports function for some telomere-like repeats in general and says nothing about the head-to-head architecture at 2q13.

The chimpanzee genome and the problem of biological similarity

Wood, Todd Charles.  Occasional Papers of the Baraminology Study Group 7:1–18, 2006.  Library key wood-2006.

camp: creationist topic: yec-argument method: review direction: complicates verified: full

Key claim: fixed chromosome differences "could have occurred after creation";  "it is difficult to imagine a scenario other than chromosomal fusion" for the 2q13 array (printed p. 7, PDF p. 9).

The primary young-earth statement of the within-kind fusion view the article tests in Part 6, cited so that the view is stated in a holder's own words before it is priced.

One pedigree we all may have come from: did Adam and Eve have the chromosome 2 fusion?

Stankiewicz, Paweł.  Molecular Cytogenetics 9:72, 2016.  https://doi.org/10.1186/s13039-016-0283-3

camp: mainstream topic: fixation method: hypothesis direction: complicates verified: full

Key claims: a single non-recurrent fusion spread through a small polygamous clan under a bottleneck;  13 families homozygous for a Robertsonian fusion;  a healthy 44-chromosome man with normal spermatogenesis.

Concedes that fixation is a real problem and answers with a hypothesis rather than a model.  Cited for the fixation problem and the Song case, not for its adaptive speculation.  CC BY 4.0.

Considered and not used

• Bergman & Tomkins 2011, part 1. Read in full. Not cited for any number; see its entry above.
• Drets & Therman 1983, the 6;19 constitutional telomere fusion IJdo cites. Not fetched. Letsolo 2010 supersedes it with 298 sequenced junctions.
• Farré et al. 2009 (Cytogenetic and Genome Research), "Interstitial telomeric sequences (ITSs) are not located at the exact evolutionary breakpoints in primates", and Richard Sternberg 2009 (Evolution News), who cites it to argue that 2q13 is the only ITS tied to a breakpoint and so is atypical. Both found in Eric's notes (/home/share/notes/eric/research/biology/chromosome 2.md) during the 2026-10-27 strengthening review and not fetched. The quoted line, that only the 2q13 head-to-head ITS "is precisely associated with an evolutionary breakpoint", bears directly on Issue 1 (is 2q13 unique?) and should be read from the paper before Issue 1 is run.
• Nothing else. The 1979 sheep study, Kasai 2000, Capper 2007 and Meyer 2012 were all obtained during verification and are cited; see their entries above.

Could not be obtained

• Yunis & Prakash 1982, full text. Closed at Science; the registered AAAS account does not open the 1982 back catalogue. Cited from the abstract as the origin of the prediction.
• Miga 2016, full text. Paid access at Oxford Academic. The abstract carries the archaic-genome result the article uses. An author copy from UCSC or a library would open it.
• Avarello 1992, full text. Springer paywall. The abstract carries the whole claim.
• Kasai et al. 2000, full text. Springer paywall; Unpaywall and Semantic Scholar both report it closed. The abstract states the 38-cosmid result in full and is now cited directly rather than through Fan.
• Bruère & Ellis 1979, full text. Closed at Oxford University Press, which now publishes the journal. The abstract carries both quotations used.
• Capper et al. 2007 was obtained in full during verification with bin/pmcget.py and is now in the source library. Meyer et al. 2012 was already in the library and is now cited directly.