Behind the branches
Curator’s Notes & FAQ
About the roster & exceptions ↗ · Explorer selection & familiar names ↗
Questions about a dinosaur, a clue or your record? Find the answers here, with the research behind our roster and tree below.
Approved exceptions
Where we amend the database
These are our approved exceptions to the captured PBDB records: names we hold out, metadata we correct, and locality assignments we reconcile. Open any entry to see the source record, our treatment, the reason and references. Expanded-roster decisions are labelled; they do not add animals to the live sample.
Clue amendments · reviewed 11 October 2026Correcting two occurrence clues; retaining four short-tree answers
Lewisuchus is Triassic. Megalosaurus uses its modern Jurassic, European identity. The four shortest answer paths pass a focused fairness check.
PBDB snapshot
Lewisuchus collection 26092 combines Longobardian and Judithian intervals, producing an implausible three-period screening range. Megalosaurus combines numerous old generic referrals across different ages and regions.
In the game
Use Triassic for Lewisuchus, and Jurassic / Europe for Megalosaurus. Keep the verified Megalosaurus type locality as confirmed; preserve other historical locality referrals as uncertain and searchable. Retain Archaeocursor, Asilisaurus, Lewisuchus and Soumyasaurus as hidden answers.
Why this choice
A collection metadata error and historical generic identifications should not control current genus clues. The four shallow genera can be isolated to four possibilities through tree feedback without clues; their standard clues then distinguish them. Unsupported family labels would give a false impression of certainty.
What it changes
Period and region clues, journal filters and locality status now agree with the reviewed profiles. Captured PBDB records are unchanged. The full follow-up records the evidence and test scope.
Classification choice · reviewed 11 October 2026Unaysauridae: our chosen published working model
Retain the established family model while documenting the newer weakly supported alternative.
In the game
Retain the existing published Unaysauridae membership of Macrocollum, Unaysaurus and Jaklapallisaurus, supported by the dedicated 2024 studies. Retain Musankwa at Massopoda and Pradhania in the existing Massospondylidae working model. The weak and partly alternative 2026 topology is disclosed but is not adopted piecemeal. Reconsider as a whole if a better-resolved analysis changes the balance of evidence.
Why this choice
The dedicated 2024 osteological studies support the existing family membership. The 2026 Musango study has weak support and alternative higher arrangements. We retain a documented playable model without treating publication recency alone as scientific certainty.
What it changes
No new genus, unsupported subgroup or isolated family expansion is added. This closes the release decision; future evidence can still justify a complete revision.
Database exceptions · approved 11 October 2026Seven exclusions, one accepted synonym and two held guesses
These decisions apply the completed working-tree audit to the active Expert roster.
In the game
Exclude Saltopus, Protoavis, Yaverlandia, Krzyzanowskisaurus, Ferganocephale, Gongbusaurus and Nanosaurus. Accept Stormbergia as a guess that solves Lesothosaurus, with one canonical Lesothosaurus discovery. Retain Serendipaceratops and Notoceratops under Genasauria as legal guesses, with independent answer holds.
Why this choice
The published revisions identify uncertain dinosaur scope, composite or indeterminate name-bearing material, and a chosen Lesothosaurus synonym treatment. Broader ancestry is retained where it is defensible. These are explicit exceptions to the captured PBDB register, not silent reassignment of fossil specimens.
What it changes
Expert now has 1,274 legal guesses and 1,236 hidden answers. Every exclusion has sources and a reconsideration condition. Existing discoveries and retired guess histories are preserved; the sample is unchanged.
Classification amendment · reviewed 11 October 2026Chilesaurus: a plant-eating theropod in our working tree
We choose a basal tetanuran placement, rather than leaving this genus attached only to Dinosauria.
PBDB snapshot
The captured PBDB record places Chilesaurus within Ornithischia. Our earlier working tree collapsed the disputed alternatives to Dinosauria. The accepted genus identity is retained.
In the game
Place Chilesaurus within Theropoda, Neotheropoda, Averostra and Tetanurae, without assigning it to a narrower family. It remains a herbivore.
Why this choice
The original description and Cau’s 2024 analysis support a tetanuran interpretation; Hendrickx’s 2026 dental study discusses this treatment alongside the competing ornithischian results. We choose this published topology to give players useful branch feedback. It is a working interpretation, not a claim that scientific disagreement has ended.
What it changes
The tree, group filter and guess feedback now use the same theropod placement. Fossil identities, localities, diet and inclusion remain separate from this ancestry choice.
Classification amendments · reviewed 11 October 2026A tree audit, with uncertainty kept visible
The audit checked all 1,281 paths in the prior roster and investigated 86 shallow paths. The approved follow-up leaves 1,274 active legal guesses.
PBDB snapshot
Original PBDB identities and captured paths remain preserved. Seven explicit identity/scope exclusions are now approved; inclusion and hidden-answer eligibility remain separate.
In the game
Ten early saurischian genera gain supported working ancestry. Four southern armoured genera share Parankylosauria. The early ornithischian grade now uses coherent shared ancestors, and the imported many-member Silesauridae group is removed under this selected hypothesis.
Why this choice
Shallow positions can be legitimate. We use published working hypotheses and disclose competing results, rather than inventing families to make every path equally deep. The full parse checks tree integrity; it is not an independent scientific certification of all genera.
What it changes
Seven uncertain identities are excluded from active guessing. Stormbergia is an accepted equivalent of Lesothosaurus. Serendipaceratops and Notoceratops remain broad guesses on answer holds. Nyasasaurus retains its unresolved administrative root and answer hold. Retired names remain readable in prior guess histories.
Expert reference · reviewed 11 October 2026Our database exceptions, in one place
1,274 legal guesses; 47 source identities excluded. Inspect every name, reason and source in the searchable Expert reference.
In the game
The original Expert curation excluded 21 further captured identities: Coloradia, Fenestrosaurus, Ovoraptor, Sangonghesaurus, Ngexisaurus, Changdusaurus, Eugongbusaurus, Himeoolithus, Sauropus, Yezosaurus, Actiosaurus, Sauraechmodon, Priodontosaurus, Ornithomimidorum, Avalonianus, Syntarsus, Ingenia, Koutalisaurus, Saurophaganax Bagaraatan and Alwalkeria. The approved working-tree follow-up excludes Saltopus, Protoavis, Yaverlandia, Krzyzanowskisaurus, Ferganocephale, Gongbusaurus and Nanosaurus.
Why this choice
The decisions distinguish unavailable or preoccupied names, a non-genus record, material outside the dinosaur scope, and exceptionally indeterminate or composite name-bearing material. They are individually sourced; disputed genera are not automatically cut. A replacement name or revised referral does not transfer every historical fossil.
What it changes
There are 47 exclusions in total. Original IDs, source records and reconsideration conditions remain visible. Expert has 1,274 legal guesses; the original sample stays separate.
Database exceptions · approved 11 October 2026Troodon, Alcovasaurus and Alwalkeria: approved release decisions
Troodon is the hidden answer for three equivalent guesses; Alcovasaurus remains a guess; Alwalkeria is excluded.
In the game
Use the published 2025 Troodon framework: Troodon, Stenonychosaurus and Latenivenatrix all solve hidden Troodon. Stenonychosaurus and Latenivenatrix are not separate hidden answers. Retain Alcovasaurus as a legal guess with an answer hold. Exclude Alwalkeria under the reviewed non-avian dinosaur scope.
Why this choice
The Troodon framework is an explicit literature-based identity choice, not a claim of universal consensus. Original source identities and fossil records remain distinct. Alcovasaurus has unresolved diagnostic and identity concerns; Alwalkeria lacks secure dinosaurian placement in the reviewed material.
What it changes
Approved by the owner on 11 October 2026. Cryptarcus and Dinevenator await a verified PBDB baseline refresh; no source IDs are invented.
Approved inclusion exception · expanded rosterDachongosaurus: unavailable informal name
An undescribed informal name, held out pending a qualifying formal description.
PBDB snapshot
The frozen PBDB snapshot selects Dachongosaurus (313245) as an accepted genus.
In the game
Exclude this identity from expanded guesses and hidden answers pending the published reconsideration condition. Preserve the source record.
Why this choice
Explicit nomen nudum for Dachungosaurus, with Dachongosaurus alternative spelling. Revision states remains undescribed although an in-situ photograph was published. A photographed skeleton alone is not an independently diagnosed genus.
What it changes
These six exclusions were approved at the historical v11 checkpoint. The current Expert reference has 1,274 selected guesses. The live sample remains 36; no fossils, ancestry or aliases are transferred.
Held records: Dachongosaurus · Zhao 1985.
Approved inclusion exception · expanded rosterOshanosaurus: unavailable informal name
An undescribed informal name, held out pending a qualifying formal description.
PBDB snapshot
The frozen PBDB snapshot selects Oshanosaurus (313242) as an accepted genus.
In the game
Exclude this identity from expanded guesses and hidden answers pending the published reconsideration condition. Preserve the source record.
Why this choice
Explicit nomen nudum. Revision states remains undescribed and appears in a faunal list. No later formal validation found. Do not assume Eshanosaurus synonymy or automatically redirect.
What it changes
These six exclusions were approved at the historical v11 checkpoint. The current Expert reference has 1,274 selected guesses. The live sample remains 36; no fossils, ancestry or aliases are transferred.
Held records: Oshanosaurus · Zhao 1985.
Approved inclusion exception · expanded rosterDamalasaurus: unavailable informal name
An undescribed informal name, held out pending a qualifying formal description.
PBDB snapshot
The frozen PBDB snapshot selects Damalasaurus (313240) as an accepted genus.
In the game
Exclude this identity from expanded guesses and hidden answers pending the published reconsideration condition. Preserve the source record.
Why this choice
Explicit nomen nudum. Historical magnus/laticostalis usages do not establish available independently diagnosed genus. Zhao1983 p302 mentions Damalosaurus only as evolutionary example. Geographic disambiguation does not repair nomenclatural availability.
What it changes
These six exclusions were approved at the historical v11 checkpoint. The current Expert reference has 1,274 selected guesses. The live sample remains 36; no fossils, ancestry or aliases are transferred.
Held records: Damalasaurus · Zhao 1985.
Approved inclusion exception · expanded rosterMicrodontosaurus: unavailable informal name
An undescribed informal name, held out pending a qualifying formal description.
PBDB snapshot
The frozen PBDB snapshot selects Microdontosaurus (65648) as an accepted genus.
In the game
Exclude this identity from expanded guesses and hidden answers pending the published reconsideration condition. Preserve the source record.
Why this choice
Explicit nomen nudum. Zhao1983 p304 provides only a brief evolutionary example. Geographic and Chaya locality data do not establish nomenclatural availability.
What it changes
These six exclusions were approved at the historical v11 checkpoint. The current Expert reference has 1,274 selected guesses. The live sample remains 36; no fossils, ancestry or aliases are transferred.
Held records: Microdontosaurus · Zhao 1983.
Approved inclusion exception · expanded rosterPonerosteus: dubious material
The naming author considered the material unidentifiable and possibly not dinosaurian.
PBDB snapshot
The frozen PBDB snapshot selects Ponerosteus (92286) as an accepted genus.
In the game
Exclude this identity from expanded guesses and hidden answers pending the published reconsideration condition. Preserve the source record.
Why this choice
Olshevsky 2000 naming-author statement calls the genus and species nomina dubia and doubtfully dinosaurian. A specimen which cannot securely establish even dinosaur affinity or an independently diagnosable genus is unsuitable as a separate playable genus.
What it changes
These six exclusions were approved at the historical v11 checkpoint. The current Expert reference has 1,274 selected guesses. The live sample remains 36; no fossils, ancestry or aliases are transferred.
Held records: Ponerosteus · Olshevsky 2000.
Approved inclusion exception · expanded rosterXuanhuasaurus: undescribed historical identity
The same original specimen was later described as Xuanhuaceratops; one animal, not two.
PBDB snapshot
The frozen PBDB snapshot selects Xuanhuasaurus (313238) as an accepted genus.
In the game
Exclude this identity from expanded guesses and hidden answers pending the published reconsideration condition. Preserve the source record.
Why this choice
Zhao et al. 2006 p467 explicitly labels the earlier Xuanhuasaurus niei nomen dubium while stating no holotype, diagnosis or description was provided. Olshevsky 1991 p159 labels Xuanhuasaurus nomen nudum. An undescribed historical name for the material later formally described as Xuanhuaceratops; retained in the history, not counted as a separate playable genus.
What it changes
These six exclusions were approved at the historical v11 checkpoint. The current Expert reference has 1,274 selected guesses. The live sample remains 36; no fossils, ancestry or aliases are transferred.
Retained candidates: Xuanhuaceratops · Zhao et al. 2006.
Held records: Xuanhuasaurus · Zhao 1985.
Approved metadata correction · expanded reviewOmeisaurus: the type is above Daanzhai
Clear an incorrect member label for one occurrence; preserve the uncertain Lower/Upper Shaximiao correlation.
PBDB snapshot
Type occurrence 474577 records Shaximiao Formation and Daanzhai Member.
In the game
In the reviewed evidence overlay, clear only that occurrence’s Daanzhai member. Keep broad Shaximiao and every other occurrence or qualified referral.
Why this choice
The detailed 1983 type-site account places the fossil 200 m above the limestone now called Daanzhai. Later Lower Shaximiao summaries and the older basal Upper interpretation disagree; neither warrants a forced member assignment.
What it changes
The raw capture stays intact. This specific error no longer creates a Daanzhai clue. The genus remains a selected guess; no live answer is added.
Approved clue restriction · expanded reviewXinjiangtitan: secure locality, disputed formation
Retain Qiketai and Qigu as interpretations of the same fossil. Use broad Jurassic and secure geography.
PBDB snapshot
Occurrence 1190808, collection 153762, assigns SSV 12001 to Qiketai.
In the game
Keep both Qiketai and the literature-backed Qigu alternative in the reviewed footprint. Use Turpan Basin, Xinjiang, China for locality help; neither disputed formation, a fine Jurassic subdivision nor a regional numerical date may decisively eliminate a guess.
Why this choice
The original description and later specimen research use Qigu; other published work argues for Qiketai. The inspected papers do not supply decisive specimen-specific mapping or dating that settles the conflict.
What it changes
Keep the guess. The province alone leaves six candidates; Turpan Basin leaves at most four, including Tienshanosaurus whose two basin-unknown referrals remain possible. The crowd issue is resolved without a discovery-date clue; normal Expert approval remains.
Approved hidden-answer hold · expanded reviewQijianglong: an uncertain horizon and period
Keep the accepted guess, retain both horizon interpretations, and hold an answer with an unqualified period clue.
PBDB snapshot
Type occurrence 1253472, collection 165720, records Suining Formation.
In the game
Preserve Suining and the reported Upper Shaximiao alternative for the same QJGPM 1001 fossil. Treat period as unresolved in conservative comparisons, so a working Jurassic label cannot wrongly eliminate it under another period clue.
Why this choice
The 2025 remapping claim relies on explicitly unpublished surveys. Other 2025 work retains Suining; regional studies disagree about Jurassic versus Cretaceous age. The 2026 correction only repairs a name in a figure, not this geological issue.
What it changes
Retain the legal guess. Hold hidden-answer release until direct type-quarry evidence or an approved uncertainty-aware period clue is available. Chongqing geography remains usable. No confirmed Cretaceous occurrence is invented.
Approved hidden-answer hold · expanded reviewZigongosaurus: a guess, but not yet an answer
Keep the accepted PBDB genus. Composite material and the doubtful species record need a specimen-based resolution.
PBDB snapshot
The accepted genus 66512 is classified in Mamenchisauridae. Its original species record 56465, recombined as Mamenchisaurus fuxiensis, is marked nomen dubium by a separate PBDB opinion.
In the game
Retain Zigongosaurus as an expanded-roster guess; hold it as a hidden answer. Do not impose a synonym, merge occurrence evidence, or silently equate museum numbers.
Why this choice
The original 1976 paper identifies specimen 02501 and prepared material from at least four individuals. The 2005 synopsis describes a composite holotype. Omeisaurus fuxiensis has its own type CV 00267. Modern papers still use Zigongosaurus separately, but do not resolve the older associations. The doubtful-species status is verified in PBDB; the full cited Wang2019 taxonomic passage remains unread. The inclusion review retains the guess because the 2005 synopsis publishes a diagnosis and the 2025 Tongnanlong paper compares named specimens. Neither resolves the mixed original associations.
What it changes
No genus is cut. Release needs a coherent independent specimen identity and supported clues. Use original number 02501, conventionally CV 02501; do not repeat CV 00261 without catalog evidence.
Approved exception · expanded rosterEucentrosaurus: one animal, two names
Replacement name for the same nominal type species already represented by Centrosaurus; no independent answer animal.
PBDB snapshot
The frozen PBDB record accepts Eucentrosaurus at genus rank.
In the game
Hold this standalone identity out of expanded guesses and answers. Retain Centrosaurus as the canonical PBDB identity. Its current reference answer eligibility is recorded on the Expert roster.
Why this choice
Replacement name for the same nominal type species already represented by Centrosaurus; no independent answer animal. Original1989 naming page unavailable; identity corroborated in two primary scholarly revisions. Do not claim a specific ICZN conservation ruling or settle authority-date discrepancy beyond verified evidence.
What it changes
That checkpoint held 1,308 candidates; the current inclusion register records later exceptions. Search aliases are recorded as future proposals, not added to the live autocomplete. The live sample remains 36 genera.
Retained candidates: Centrosaurus · Lambe 1905.
Held records: Eucentrosaurus · Chure and McIntosh 1989.
Approved exception · expanded rosterProiguanodon: historical identity under review
Original source uses this as a grouping within Iguanodon, not a separate described genus. It includes more than one later genus, so a single genus alias would be misleading.
PBDB snapshot
The frozen PBDB record accepts Proiguanodon at genus rank.
In the game
Hold this standalone identity out of expanded guesses and answers. Do not force a canonical replacement or transfer fossil records.
Why this choice
Original source uses this as a grouping within Iguanodon, not a separate described genus. It includes more than one later genus, so a single genus alias would be misleading. Do not assign Bernissart because it occurs in the paper title. No one canonical target: constituent species are now associated with Cumnoria52998 and Barilium167439; these modern mappings are context, not an implemented alias.
What it changes
That checkpoint held 1,308 candidates; the current inclusion register records later exceptions. Search aliases are recorded as future proposals, not added to the live autocomplete. The live sample remains 36 genera.
Held records: Proiguanodon · Van Den Broeck 1900.
Approved exception · expanded rosterProyandusaurus: one animal, two names
Both publications expressly identify the same nominal Yandusaurus multidens species; this is not evidence for a second fossil animal.
PBDB snapshot
The frozen PBDB record accepts Proyandusaurus at genus rank.
In the game
Hold this standalone identity out of expanded guesses and answers. Retain Hexinlusaurus as the canonical PBDB identity. Its current reference answer eligibility is recorded on the Expert roster.
Why this choice
Both publications expressly identify the same nominal Yandusaurus multidens species; this is not evidence for a second fossil animal. 2005 full original not obtained; same-species naming statement is explicit in the primary abstract. No locality transfer needed for a held duplicate.
What it changes
That checkpoint held 1,308 candidates; the current inclusion register records later exceptions. Search aliases are recorded as future proposals, not added to the live autocomplete. The live sample remains 36 genera.
Retained candidates: Hexinlusaurus · Barrett et al. 2005.
Held records: Proyandusaurus · Knoll 1999.
Approved exception · expanded rosterNectosaurus: one animal, two names
This is the1910 dinosaur usage for Brown’s navajovius skull, not marine Nectosaurus Merriam1905. PBDB itself records the usage as a recombination of the accepted Kritosaurus navajovius species.
PBDB snapshot
The frozen PBDB record accepts Nectosaurus at genus rank.
In the game
Hold this standalone identity out of expanded guesses and answers. Retain Kritosaurus as the canonical PBDB identity. Its current reference answer eligibility is recorded on the Expert roster.
Why this choice
This is the1910 dinosaur usage for Brown’s navajovius skull, not marine Nectosaurus Merriam1905. PBDB itself records the usage as a recombination of the accepted Kritosaurus navajovius species. Brown1910 full original not obtained in this bounded run. Correct original Versluys dinosaur usage was read, and exact identity bridge comes from fresh PBDB species opinions rather than secondary genus lists. Do not merge into the marine reptile genus or claim publication-priority details not independently established.
What it changes
That checkpoint held 1,308 candidates; the current inclusion register records later exceptions. Search aliases are recorded as future proposals, not added to the live autocomplete. The live sample remains 36 genera.
Retained candidates: Kritosaurus · Brown 1910.
Held records: Nectosaurus · Versluys 1910.
Approved exception · expanded rosterMicrosaurops: historical identity under review
Published revision1991 treats it as a spelling/replacement artifact under Microcoelus, not an independently described animal.
PBDB snapshot
The frozen PBDB record accepts Microsaurops at genus rank.
In the game
Hold this standalone identity out of expanded guesses and answers. Do not force a canonical replacement or transfer fossil records.
Why this choice
Published revision1991 treats it as a spelling/replacement artifact under Microcoelus, not an independently described animal. Original Kuhn entry not read; PBDB1964 versus revision1963 discrepancy disclosed. No formal Saltasaurus synonym asserted and no new genus or tree forced.
What it changes
That checkpoint held 1,308 candidates; the current inclusion register records later exceptions. Search aliases are recorded as future proposals, not added to the live autocomplete. The live sample remains 36 genera.
Held records: Microsaurops · Kuhn 1964.
Approved evidence overlay · expanded-roster auditLocality records: formations, groups and uncertain sites
Correct exact records, keep uncertain alternatives and include every documented occurrence.
PBDB snapshot
Some captured rows use Upper or 5 as formation names, omit a formation, or retain competing stratigraphic assignments.
In the game
Clear misleading Hekou formation labels; add supported Yanan and Nanxiong assignments to exact records. Add missing positive locality profiles and preserve qualified alternatives. These are audit inputs, not newly approved daily answers.
Why this choice
A group is not a formation, a database collection is not necessarily a quarry, and a type locality does not define the entire genus range. Real five-genus assemblages remain crowded.
What it changes
All 30 remaining proposals have a recorded decision. The detailed report lists corrections, retained uncertainty, source access and before/after comparisons.
Approved rank exception · expanded rosterEosauropoda: a group entered as a genus
Excluded from guesses and answers because the original publication describes a group.
PBDB snapshot
PBDB 370667 accepts Eosauropoda Bonaparte 1986 at genus rank, with no linked occurrences in our capture.
In the game
Exclude this identity from the expanded guess roster and hidden answers. The actual dinosaur genera within our scope remain eligible for the usual review.
Why this choice
Bonaparte’s cited publication discusses Eosauropoda as a grouping containing named genera, including Cetiosaurus, Datousaurus and Shunosaurus. This is a verified source-rank mismatch.
What it changes
Approved on 10 October 2026. The original database record is preserved; this exclusion introduces no ancestry change or new group in the working tree. The live sample remains 36 genera.
Held records: Eosauropoda · Bonaparte 1986.
Approved replacement-name exception · expanded rosterNemegtia becomes Nemegtomaia
Use Nemegtomaia; the preoccupied former name is not a separate playable genus.
PBDB snapshot
The snapshot separately accepts Nemegtia (68043, Lü et al. 2004) and Nemegtomaia (95490, Lü et al. 2005), with different source placements. Nemegtia has no linked occurrences; the original type occurrence is already accepted under Nemegtomaia.
In the game
Retain Nemegtomaia as the single eligible genus identity and exclude Nemegtia as an independent guess or hidden answer. Both original database records remain preserved for inspection.
Why this choice
The authors formally issued Nemegtomaia as the replacement name in 2005 because Nemegtia was already used for an ostracod. This is a documented naming change for the same dinosaur.
What it changes
Approved on 10 October 2026. The expanded selected roster now contains 1,314 genera. All three canonical occurrence records are retained, and literature supports both Nemegt and Baruungoyot formations. The second new specimen in the 2012 paper has inconsistent horizon wording; that uncertainty remains. No fossils or ancestry are transferred, and the live sample remains 36 genera.
Retained candidates: Nemegtomaia · Lü et al. 2005.
Held records: Nemegtia · Lü et al. 2004.
Approved locality supplement · expanded reviewAdamantisaurus: restoring a documented locality
Add the published Brazilian type locality; keep the exact age and revised formation unresolved.
PBDB snapshot
The captured Adamantisaurus record (70544) is accepted but has no linked occurrences. That empty set is preserved as source data.
In the game
The locality review now includes positive literature evidence from near Flórida Paulista, São Paulo, Brazil, in the broadly defined Adamantina Formation of the Bauru Group. The genus remains included.
Why this choice
The original description names the holotype series MUGEO 1282, 1289 and 1295. Langer et al. 2022 corroborates the locality and lists uncertain Vale do Rio do Peixe/Presidente Prudente placement. We retain Adamantina broadly. The 2025 summary has a catalogue-number inconsistency, so we preserve the original specimen identifiers. Associated femur and teeth are not automatically assigned to this genus.
What it changes
The known type locality supplements the comparison catalogue without fabricating a PBDB occurrence. Late Cretaceous is the supported broad age; no exact stage or date is assigned. Positive locality information is not proof of absence elsewhere. Standard-clue source screening, ancestry, diet and hidden-answer eligibility remain under their existing review rules. The expanded roster remains 1,314 genera, and the live sample remains 36.
Retained candidates: Adamantisaurus · Santucci and Bertini 2006.
Name resolution · expanded datasetOne Saurornithoides identity
Use Osborn’s troodontid record; hold the separate Nopcsa identity.
PBDB snapshot
The captured database has two accepted records named Saurornithoides: Osborn 1924 (38573), under Troodontidae, with eight linked occurrences; and Nopcsa 1928 (69227), under Compsognathinae, with none.
In the game
Use the Osborn 1924 record (38573) for Saurornithoides in the expanded roster. Hold the Nopcsa 1928 record (69227) out for nomenclatural review. Neither is in the live 36-genus sample.
Why this choice
The published taxonomic review identifies the familiar Mongolian troodontid as Saurornithoides Osborn 1924. A typed genus needs one explicit identity so its relationships never depend on which database row happens to come first.
What it changes
Saurornithoides will follow the retained Osborn record. The two records and their fossil evidence are not merged. Current reference eligibility is recorded on the Expert roster; the live sample is unchanged.
Retained candidates: Saurornithoides · Osborn 1924.
Held records: Saurornithoides · Nopcsa 1928.
Metadata correction · expanded datasetTwo fossil genera marked as living
Correct the extinct-status flags for Kansaignathus and Kurupi.
PBDB snapshot
Kansaignathus (444043) and Kurupi (443993) are accepted non-avian theropod genera in the captured hierarchy, but both carry an extant flag. Their linked fossil occurrences are Cretaceous.
In the game
Retain both for the expanded roster and locally mark them extinct. Their PBDB identities and recorded ancestry stay intact. They are not yet part of the live 36-genus sample.
Why this choice
Their original descriptions identify fossil dinosaurs: Kansaignathus from the Santonian of Tajikistan and Kurupi from the Maastrichtian of Brazil. The living flag contradicts the described evidence.
What it changes
A metadata error will not remove either genus from the future roster. This corrects a data field; it does not settle finer relationships or approve daily-answer eligibility. Bird exclusions still use the named database subtrees.
Retained candidates: Kansaignathus · Averianov and Lopatin 2021, Kurupi · Iori et al. 2021.
Temporary record hold · expanded datasetAncistrodon: a tangled name record
Hold a record whose name, classification and linked fossil evidence conflict.
PBDB snapshot
Ancistrodon (154360), attributed to Huxley 1865, is attached directly to Dinosauria. Its only captured occurrence is Ancistrodon libycus from the Ypresian of Saudi Arabia (occurrence 434618, reference 11298).
In the game
Hold this record out of the expanded guess and answer pools until its identity and linked evidence can be reconciled. Keep the raw record available for inspection.
Why this choice
Priem’s fossil study places Ancistrodon libycus among fish. A modern review identifies Huxley’s similarly named Ankistrodon indicus as an Early Triassic proterosuchid archosauriform. These conflicting identities are evidence of a likely name-linking problem, rather than a dinosaur surviving into the Eocene.
What it changes
No invented dinosaur period, diet, region or ancestry is assigned. We do not turn this record into an Ankistrodon alias. The precise database correction remains unresolved; this is a temporary hold on a conflicting record, not a general rule excluding scientifically disputed genera.
Held records: Ancistrodon · Huxley 1865.
Approved identity exceptionShuvuuia: one genus, two source spellings
Use Shuvuuia; exclude the duplicate Shuuvuia spelling identity.
PBDB snapshot
Shuvuuia (66043) and Shuuvuia (307772) are both self-accepted in our snapshot and the fresh PBDB identity check. Both cite Chiappe et al. 1998, reference 13618. Shuvuuia has three linked occurrences and its type species; Shuuvuia has none. Opinions on the populated record mark its spelling as a correction.
In the game
Use only Shuvuuia (66043) in the expanded review roster. Hold Shuuvuia (307772) out as a duplicate spelling identity; it is not a separate guess or hidden answer. Both original source records remain available. The 36-genus live sample is unchanged.
Why this choice
The shared original description, spelling-correction opinions and primary systematic treatment support one genus. This explicit exception to PBDB accepted identities was approved on 10 October 2026.
What it changes
The updated Baruungoyot check leaves four after an incorrect guess of one of its five documented genera, or five after an outside guess. No occurrence or ancestry evidence was merged. Current reference eligibility is recorded on the Expert roster.
Retained candidates: Shuvuuia · Chiappe et al. 1998.
Held records: Shuuvuia · Chiappe et al. 1998.
Approved identity exceptionHylaeosaurus: one dinosaur, an extra spelling
Use Hylaeosaurus; exclude the duplicate Hylosaurus spelling identity.
PBDB snapshot
Hylaeosaurus (38822) and Hylosaurus (66560) both remain accepted. Both cite Mantell 1833 and reference 23493. The empty Hylosaurus record has one classification opinion, explicitly marked misspelling.
In the game
Use only Hylaeosaurus; hold the extra Hylosaurus spelling record out of the expanded roster. Preserve the original records and their separate fossil links.
Why this choice
The shared original description and explicit misspelling flag support a spelling correction. Fitzinger’s later Hylosaurus mantelli is a distinct historical usage treated as a name lacking a sufficient description in Raven’s review.
What it changes
The extra unknown candidate no longer inflates locality comparisons. This changes the review queue, without enabling new hidden answers or merging fossil evidence.
Retained candidates: Hylaeosaurus · Mantell 1833.
Held records: Hylosaurus · Mantell 1833.
Approved hold · synonymy review remains openHoplosaurus: untangling two historical usages
Hold Hoplosaurus as a separate genus; retain Struthiosaurus.
PBDB snapshot
Hoplosaurus Seeley 1881 (53046) remains accepted, with no linked occurrences or child species. Its selected opinion is Kuhn 1964; older opinions instead link the name to the sauropod Pelorosaurus.
In the game
Retain Struthiosaurus and hold Hoplosaurus out as an independent genus. Do not transfer fossils or ancestry from the held record.
Why this choice
The Dinosauria distribution chapter includes the Austrian Hoplosaurus ischyrus within Struthiosaurus austriacus. Sauropod literature separately documents Hoplosaurus as a misspelling of Oplosaurus. The database opinions appear to mix these usages.
What it changes
The extra unknown candidate no longer inflates locality comparisons. The full 2001 reappraisal and exact formal synonymy still need checking; this is a documented hold, not a claim of completed synonymisation.
Retained candidates: Struthiosaurus · Bunzel 1870.
Held records: Hoplosaurus · Seeley 1881.
Expanded-roster evidenceGastonia: untangling the rock layers
Reconcile specific Cedar Mountain horizons while preserving uncertain referrals.
PBDB snapshot
The snapshot labels Animantarx’s Carol Site and Peloroplites’ Price River II record Ruby Ranch. Gastonia is recorded in Yellow Cat and Poison Strip; a Kansas Silvisaurus record lacks formation information.
In the game
The review corrects the exact Animantarx row to Mussentuchit using its original type description. Peloroplites retains both supported horizons because newer literature still reports Ruby Ranch material. Lorrie’s Site likewise retains Poison Strip and Ruby Ranch alternatives.
Why this choice
Different or incomplete record labels must not manufacture a five-genus Yellow Cat overlap. Formation summaries describe known records rather than additional fossils with unknown horizons.
What it changes
With the approved Hylosaurus and Hoplosaurus holds applied, Poison Strip leaves 3–4 candidates after tested wrong guesses and Yellow Cat leaves 1–2. The unknown Silvisaurus record remains possible, and qualified Poison Strip referrals are preserved. This review does not enable a new playable answer.
Frequently asked questions
Why is a dinosaur missing from the game?
Explorer is a smaller selection of familiar genera. Expert follows our captured Paleobiology Database (PBDB) roster, with documented exceptions for records such as synonyms or dubious names. New descriptions may await a verified database update.
Check the Explorer selection or search the full roster and exclusions. The database exceptions above explain our departures.
Why can I guess a dinosaur that is never an answer?
Some included genera have an answer hold while their identity or clues remain unsuitable for a fair puzzle. They remain legal guesses. Their journal entries and the full roster explain each hold.
Which sources set the tree and clues?
PBDB is our starting point for which genera are included. Reviewed scientific literature informs relationships and clues, with amendments recorded below. Diet labels use the best-supported broad interpretation; some rely on close relatives. These are useful game categories, not claims that every detail is certain.
Region and dig-site records describe known finds, not an exclusive range. Images illustrate a genus or a named member of a group; they do not set the clues.
How do clues and free field reports work?
Reveal Diet, Period or Region in any order. Each reveal counts as one clue. A free field report may help when six or more possible answers cannot be usefully separated by tree feedback and standard clues. Free reports do not add to your clue count.
What changes between Explorer, Expert, Daily and Practice?
Explorer uses familiar genera; Expert uses the full reviewed roster. Both contribute to the same journal, which always shows every genus. The discovery border records the roster used. Successful Daily expeditions also earn a dated stamp; Practice discoveries do not. Each roster has its own Daily puzzle, resetting at midnight UTC.
How do I keep my discoveries?
Guest progress is saved in this browser. Sign in to save your journal automatically to your account and resume on other devices. Clearing browser data removes the device’s copy; progress already saved to your account remains available. The game’s footer has more information under Data & storage.
2 notesWhat belongs in the journal
We use PBDB accepted genera within our stated scope. Explicit name holds and evidence-backed metadata corrections are recorded here by database identity.
Expert reference · reviewed 11 October 2026A legal guess is not always a fair answer
1,236 eligible reference profiles; 38 conservative answer holds.
In the game
Keep defensible PBDB guesses even when their evidence does not support an independent hidden answer. Holds cover subjective duplicate identities, unresolved material or availability, and missing dependable clues.
Why this choice
An answer hold is not a blanket declaration that a genus is invalid. Fourteen Pakistani names remain legal guesses while independent availability and specimen associations are unresolved; they are not dismissed as informal names despite later descriptions.
What it changes
Expert mode is available for daily and practice games with 1,236 eligible answers. The 38 held profiles remain legal guesses. Approved identity decisions are linked below; the original sample and its saved rounds are separate.
Roster policy · expanded datasetThirteen early-dinosaur candidates retained
PBDB snapshot
The captured PBDB hierarchy accepts these 13 genera within the Parapredentata branch of Ornithischia, outside Saphornithischia. They are outside the excluded Aves and Avialae subtrees.
In the game
Retain them for the expanded roster under the same accepted-genus rule used elsewhere. They are included in Expert; the original sample stays separate.
Why this choice
A disputed placement elsewhere in the literature is not, on its own, grounds to reject an accepted PBDB genus. We keep the source identity and placement, with any disagreement explained in its journal entry.
What it changes
Inclusion is approved. Daily-answer eligibility still depends on reviewed placement and usable clue evidence. The current Expert reference records supported working-tree amendments separately.
Retained candidates: Amanasaurus, Asilisaurus, Diodorus, Eucoelophysis, Gamatavus, Kwanasaurus, Lewisuchus, Lutungutali, Pisanosaurus, Sacisaurus, Silesaurus, Soumyasaurus, Technosaurus.
6 notesConnecting the branches
These five amendments change the recorded ancestry used for guess feedback. They are selected, sourced relationships in the game's working tree.
Expert reference · reviewed 11 October 2026The full working tree
Incorrect parents corrected; disputed fine placements broadened to supported common ancestors.
In the game
The Expert reference records every raw and working lineage. Examples include Volgatitan within Titanosauria, Lingwulong within Dicraeosauridae, Geminiraptor within Therizinosauria after the September 2026 revision, and a basal Hadrosaurus outside Brachylophosaurini.
Why this choice
A consistent tree needs one parent for each displayed group. The chosen basal-Hadrosaurus model uses Saurolophinae for the other crestless branches; Hadrosaurinae and Saurolophinae are not declared unconditional synonyms. Conflicting published small-theropod and South American hadrosaur placements use broader supported parents.
What it changes
664 selected working paths differ from the raw capture, including the declared backbone and shared group rewrites. This is not a count of database errors or independent revisions. Sources and access limits are recorded per amendment.
Working-tree amendmentA shared Saurischia branch
PBDB snapshot
In the captured hierarchy, Theropoda connects directly to Dinosauria. Sauropodomorpha connects through Saurischia.
In the game
Saurischia is inserted above every represented theropod, joining theropods and sauropodomorphs within that group.
Why this choice
The prototype selects the traditional Saurischia arrangement. Langer and colleagues recovered this grouping in their reanalysis; Baron and colleagues proposed an alternative. This is a declared working hypothesis, not a settled boundary between the major branches.
What it changes
A Tyrannosaurus guess against Diplodocus reports Saurischia, rather than the raw export's Dinosauria.
Current entries: Albertosaurus, Allosaurus, Baryonyx, Carnotaurus, Ceratosaurus, Coelophysis, Deinonychus, Dilophosaurus, Gallimimus, Gorgosaurus, Microraptor, Oviraptor, Spinosaurus, Tarbosaurus, Tyrannosaurus, Utahraptor, Velociraptor.
Working-tree amendmentJoining the ornithischian branches
PBDB snapshot
The sampled Neornithischia paths lack Genasauria, while the sampled Thyreophora paths include it.
In the game
Genasauria is inserted above Neornithischia wherever it is missing.
Why this choice
The game uses Genasauria to connect the represented neornithischian and thyreophoran branches, following the named relationships in Madzia and colleagues' reference classification.
What it changes
An Iguanodon guess against Stegosaurus reports Genasauria, rather than Ornithischia in the raw export.
Current entries: Edmontosaurus, Iguanodon, Maiasaura, Pachycephalosaurus, Parasaurolophus, Protoceratops, Psittacosaurus, Styracosaurus, Triceratops.
Working-tree amendmentGallimimus and its shared ancestry
PBDB snapshot
Gallimimus connects through Ornithomimidae and Ornithomimosauria directly to Theropoda in this snapshot.
In the game
Its path includes Neotheropoda, Averostra, Tetanurae, Coelurosauria and Maniraptoriformes between Theropoda and Ornithomimosauria. The Saurischia amendment also applies.
Why this choice
Published ornithomimosaur analyses place this branch within the broader coelurosaur and maniraptoriform relationships. The additional groups let those shared relationships appear in the working tree.
What it changes
Gallimimus and Velociraptor share Maniraptoriformes in the game; the raw export gives Theropoda. This amendment currently covers the sample's only ornithomimosaur. It must be applied consistently across that branch before the roster expands.
Current entries: Gallimimus.
Working-tree amendmentConnecting Maniraptora
PBDB snapshot
The represented Maniraptora paths skip Maniraptoriformes.
In the game
Maniraptoriformes is inserted immediately above Maniraptora where missing.
Why this choice
The working classification connects maniraptorans and ornithomimosaurs through this shared named group, using the same reference framework as the Gallimimus amendment.
What it changes
The represented raptors and Oviraptor can share Maniraptoriformes with Gallimimus. Adding it to Gallimimus alone would not establish that shared group.
Current entries: Deinonychus, Microraptor, Oviraptor, Utahraptor, Velociraptor.
Working-tree amendmentThe horned-dinosaur connection
PBDB snapshot
The sampled ceratopsids skip Neoceratopsia and Coronosauria, while Protoceratops passes through those groups.
In the game
Triceratops and Styracosaurus connect through Neoceratopsia and Coronosauria before Ceratopsidae.
Why this choice
Farke and colleagues define Coronosauria around the shared ancestry of Protoceratops and Triceratops. These names connect the represented branches consistently in the working tree.
What it changes
Protoceratops and either sampled ceratopsid share Coronosauria in the game, rather than Ceratopsia in the raw export.
Current entries: Styracosaurus, Triceratops.
2 notesWhich labels make the map
The prototype has two kinds of simplification. Only one is purely visual.
Gameplay-tree simplificationA shorter underlying classification
PBDB snapshot
Raw genus lineages contain additional intermediate groups, including Deinonychosauria, Pyrodontia and Neostegosauria.
In the game
The builder retains a selected set of group names and omits the others from the dataset used for feedback. The separate lineage ledger lists every omitted group for each genus.
Why this choice
The small prototype uses fewer labels to keep the map manageable. This is a game-design choice, not a scientific rejection of the omitted groups.
What it changes
Stegosaurus and Kentrosaurus share Stegosauridae in the game instead of the raw hierarchy's more specific Neostegosauria. The retained group list needs review before expansion.
Display onlyFewer labels on screen
In the game
The map's Fewer labels / All labels control hides or reveals some intermediate labels already present in the working dataset.
Why this choice
This makes a long map easier to explore, especially on a phone.
What it changes
Switching this view does not change guess feedback. All labels means all groups in the game's dataset, not every group in the raw PBDB export.
14 notesReading the evidence
PBDB supplies the guess-roster baseline, with prominent documented inclusion exceptions. Relationships and clues are separately curated using valid, relevant literature.
Clue interpretationDiet categories and uncertainty
Plain labels in the game; the evidence and interpretation here.
In the game
Herbivore, Carnivore and Omnivore are the game’s working diet categories. Clues and journal entries use the same plain labels.
Why this choice
Fossil diets are not always known with certainty. Some interpretations use direct evidence; others use anatomy and close relatives. The reviewed profiles retain their uncertainty flags, compatible categories and references. Removing “likely” from a player-facing label does not strengthen the scientific evidence or narrow the possible answers.
Expert reference · reviewed 11 October 2026Clues with their evidence attached
Diet inferences labelled; all recorded competing localities retained.
In the game
Reviewed descriptions and supported relatives fill broad diet, period and region gaps. Uncertain plant/mixed feeding retains both compatible interpretations. Ten unresolved diet profiles stay out of the answer pool.
Why this choice
A formation clue is a positive found-in statement, not the only place a genus lived. Comparisons retain recorded non-primary, qualified and unknown localities. The reference includes possible Conchoraptor material from Nemegt, conflicting Peloroplites member labels and the newly sourced Minotaurasaurus and Dzharacursor referrals.
What it changes
Every eligible profile has a conservative blind-sibling crowding certificate. 442 profiles use sourced free locality reports; no screened crowd remains above five. The check assumes standard clues and the direct branch have been reached, and does not prove an optimal fifteen-guess strategy from the root.
Expanded-roster curation registerThe earlier source-coverage checkpoint
Preserved v11 register: 1,302 selected genera before the final evidence integration.
In the game
The roster desk keeps guess inclusion, scientific evidence, current prototype clues and Expert release approval separate. Search all selected genera or focus on missing inputs, named scientific follow-ups, aliases and fairness checks.
Why this choice
Later reviews have filled some original database gaps. A single register recognises those additions without treating a positive locality or a passing candidate count as a complete answer profile.
What it changes
This historical register preserves the original missing-input counts and review leads. Use the Expert reference for current 1,274 guesses, 1,236 eligible profiles and 38 answer holds.
Source policyNames from PBDB; evidence from the literature
PBDB snapshot
The dated PBDB accepted-genus roster, within our published non-avian scope and explicit identity holds, is the reference for which genera can be guessed.
In the game
For relationships, diet, age, geography and locality information, use the latest valid, relevant literature as needed. Compare evidential support and retain unresolved uncertainty; a newer publication does not automatically settle a disputed claim.
Why this choice
Players should be able to verify whether a name is a viable guess while benefiting from supported scientific corrections to other information.
What it changes
Material departures are referenced and explained here, with original evidence preserved. This policy does not automatically approve a new tree, revise every clue or expand the current 36-genus prototype.
Clue curationBroad clues from museum profiles
PBDB snapshot
PBDB diets can be inherited from a larger group. Occurrences can contain uncertain or historical identifications and broad age bounds.
In the game
The current clues use broad diet and period categories from Natural History Museum profiles where an extract is preserved. The builder takes the first matching Triassic, Jurassic or Cretaceous word, apart from the Coelophysis override. Nine profiles are entered manually; their preserved evidence still needs review.
Why this choice
The prototype uses readable broad categories rather than presenting raw occurrence metadata as a vetted range. The first-period rule and uneven evidence coverage are limitations of this sample, not the intended full-roster method.
What it changes
Clues and journal filters use the same stored values. Some current period clues may omit part of a genus's range. Expand the entry below to inspect all nine manual profiles.
Evidence review pending. These nine entries have hand-entered facts without a matching preserved museum extract. Their displayed values are listed here for inspection.
- Brontosaurus
- Herbivore · Jurassic · Americas
- Camarasaurus
- Herbivore · Jurassic · Americas
- Ceratosaurus
- Carnivore · Jurassic · Americas / Europe
- Diplodocus
- Herbivore · Jurassic · Americas
- Edmontosaurus
- Herbivore · Cretaceous · Americas
- Oviraptor
- Likely Omnivore · Cretaceous · Asia
- Styracosaurus
- Herbivore · Cretaceous · Americas
- Tarbosaurus
- Carnivore · Cretaceous · Asia
- Triceratops
- Herbivore · Cretaceous · Americas
Clue override · review pendingCoelophysis spans two period clues
In the game
Coelophysis is explicitly assigned both Triassic and Jurassic. Its Africa and Americas regions come from the museum profile's broader historical treatment.
Why this choice
The prototype follows that broad profile instead of the builder's single-period rule. PBDB accepts Coelophysis and Megapnosaurus separately, so the species assignments behind these clues need reconciliation.
What it changes
Either period filter retains Coelophysis. We have not changed these clues or declared their historical assignments resolved.
Current entries: Coelophysis.
Clue overridesRegional additions and a fixed locality value
In the game
Europe is explicitly added to Allosaurus and Stegosaurus beyond the museum's abbreviated USA profiles. Edmontosaurus is explicitly set to Americas; the same value also appears in its manual profile.
Why this choice
The European additions follow published assignments to those genera. Edmontosaurus's repeated assignment preserves the prototype value; it is not an additional scientific amendment.
What it changes
Europe appears in those two genera's clues and filters. Edmontosaurus remains Americas. These lists describe the current evidence choices, not exhaustive geographic ranges.
Current entries: Allosaurus, Stegosaurus, Edmontosaurus.
Clue wordingThree diets carry a qualifier
In the game
Oviraptor and Gallimimus display Likely Omnivore; Pachycephalosaurus displays Likely Herbivore. These uncertainty flags are explicit prototype choices.
Why this choice
Diet is an interpretation. The wording acknowledges uncertainty, but the audit found that the qualifier itself can distinguish an answer in a small roster. A consistent full-roster evidence standard is still needed.
What it changes
The visible clue includes Likely. The diet filter uses the broad category without a separate uncertainty filter.
Current entries: Oviraptor, Gallimimus, Pachycephalosaurus.
Clue simplificationFour modern discovery regions
In the game
Country names are grouped into Americas, Europe, Asia and Africa. The prototype maps Russia to Asia and does not support Australia or Antarctica.
Why this choice
The initial sample uses broad modern discovery regions for readable clues. It does not reconstruct prehistoric continents. Transcontinental locations need locality-level review before expansion.
What it changes
A region filter uses this stored grouping. A region absent from the prototype is not proof that the genus never occurred there.
Search interpretationRecognised names lead to one genus
In the game
T. rex, T rex, Tyrannosaurus rex and Trex resolve to Tyrannosaurus. Cryptovolans resolves to Microraptor. These are the prototype's only extra search names.
Why this choice
Common input forms and this recorded synonym help players reach a canonical genus. They do not create additional answers or discoveries.
What it changes
Testing either recognised name spends a guess on the same genus. The broader synonym list remains unimplemented; disputed accepted genera have not been merged.
Current entries: Tyrannosaurus, Microraptor.
Fairness rule · automatic supportHelp when a branch becomes a guessing crowd
PBDB snapshot
Some accepted genera share the same direct parent in the working classification. The database does not always supply finer relationships that can distinguish them.
In the game
When at least six unguessed candidates remain at a confirmed direct-parent branch and all three standard clues are identical, automatic support supplies extra referenced evidence. It must leave at most five candidates and does not count as a clue used. Five or fewer indistinguishable candidates are left for the player to test. Reports remain in the expedition history.
Why this choice
Unresolved relationships should not force players to spend optional clues on evidence needed to make the puzzle reasonable. Broad branches that still have useful splits or standard clues do not trigger help. We do not invent scientific facts or create new branches for game balance.
What it changes
The current 36-genus sample has no qualifying crowd. Twenty-seven preserved museum locality listings are visible and searchable, explicitly as abbreviated listings rather than complete geographic ranges. Current expanded-game eligibility and locality support are recorded in the Expert reference. The separate Branch Survey and Locality Survey cover the full source roster without expanding live answers.
Evidence policy · locality reviewOne named place, every matching occurrence
PBDB snapshot
Collection records link genus occurrences to geographic and stratigraphic context. They can omit details, preserve historical referrals or contain multiple records from the same physical site.
In the game
Future found-in clues must retain every plausible genus with any documented or credible uncertain occurrence in the named place. A preferred or type locality chooses the wording; it never makes another known occurrence disappear. The current museum listing field remains explicitly an archived profile listing.
Why this choice
A player should not have to guess which locality we chose to display for another genus. Type locality, first discovery, geological group, formation, member and physical dig site are different claims. Missing coverage cannot silently exclude an answer.
What it changes
All 108 large source-parent groups and all 63 matching standard-clue pools have a reproducible locality screen. Cited literature overlays preserve additional overlaps and flag disputed referrals and unit names. Potential splits remain evidence-review leads; no new hidden answers or classifications are enabled.
Research check · expanded datasetTesting clues against every remaining candidate
PBDB snapshot
The original locality screen compares direct source siblings. Descendants and records missing standard facts can remain viable guesses outside a particular shortlist.
In the game
A separate validation checks all 108 large groups and 63 standard-clue pools against wider source candidates, preserving non-primary and uncertain locality evidence and applying each actual wrong guess.
Why this choice
Five matches can become four after guessing one of them, while an outside guess may leave all five. A clue must also retain every relevant recorded match, wherever its preferred locality lies.
What it changes
Historical v10 checkpoint: The full validation covers 108 large source-parent groups and 63 standard-clue pools, retaining all 1,308 selected source candidates. Under the current five-candidate limit, 473 of 484 chosen locality comparisons are already within five; 11 need useful tree navigation. Every supported target reaches five or fewer in the demonstrated strategy. Missing facts and incomplete locality footprints still need scientific review. The current v13 Expert reference and new conservative certificates are on the Expert roster.
Research decision · expanded rosterUseful branches before extra help
PBDB snapshot
A source-tree crowd can contain several usable branches or a flat set of direct siblings. Counting formation members does not distinguish these cases.
In the game
For the expanded-roster review, we test each remaining candidate state against every legal source genus as a next guess. The current rule allows five indistinguishable candidates. Six or more need free referenced evidence when no informative tree split or standard clue remains. The live sample has no qualifying crowd.
Why this choice
All competitors matter, including those absent from the original proposal list. At Oldman, all six candidates are included and a useful next guess can leave at most three. The five Baruungoyot alvarezsaurids cannot be separated by current source-tree feedback alone.
What it changes
Historical geographic checkpoint: The earlier country-only Daye/Chaya comparison produced a possible 19-mamenchisaurid branch through unknown group fields. The follow-up reviewed all 19 genera and 32 captured occurrences. Explicit Tibet/Xizang support leaves two candidates in each target comparison; finer unit + province wording leaves one. Microdontosaurus uses Chaya, not Daye Group. The review preserves qualified and non-primary records and invents no group assignments. That was an evidence-review checkpoint; current Expert eligibility is recorded in the full reference. Baruungoyot needs no further discovery-date clue solely for crowd size. Damalasaurus and Microdontosaurus are now excluded for unavailable naming, independently of these count results.
Looking for a particular dinosaur?
The separate lineage ledger lets you search by genus or group and inspect its recorded relationships and clues. Shared amendments are explained once in these notes.
Still on the curator’s desk
Expert has 1,236 eligible answers and 38 answer holds, each with a recorded reason. The owner approved the Troodon framework, Alcovasaurus hold and Alwalkeria exclusion. Newly described Cryptarcus and Dinevenator await a verified PBDB refresh.
Illustrating the journal
A group picture shows a named member of that branch, rather than an ancestor or the mystery answer. Genus portraits show the named animal or its fossil material. Artist’s reconstructions include inferred colours and soft tissues; skeleton mounts may contain casts and reconstructed parts.
Every image links to its creator, original file and reuse licence. We keep resized copies for dependable loading. The field notes link to their sources; neither pictures nor Wikipedia summaries set the game’s clues or working tree.