In this Late Carboniferous estuarine lagoon, about 310–300 million years ago, small Adelophthalmus eurypterids—among the last and most adaptable “sea scorpions”—move across organic-rich mud at the edge of a coal-swamp forest. Around them lie Edmondia bivalves and ostracod-rich microbial mats, while the shoreline rises into peat-darkened lowlands crowded with lycopsids, Calamites, tree ferns, and seed ferns typical of the equatorial wetlands of Euramerica. The scene captures a quiet brackish habitat where terrestrial coal-forming swamps and shallow coastal waters met in one of Earth’s most distinctive tropical ecosystems.
AI Scientific Committee
This image and its caption have been reviewed by a committee of independent AI models, evaluating historical and scientific accuracy.
Claude
Image:
Adjust
Caption:
Approved
Mar 27, 2026
The image presents a broadly convincing Late Carboniferous estuarine scene with appropriate dark organic-rich sediments, quiet brackish water, and a forested backdrop. The overall mood and setting are well-conceived. However, several issues warrant attention. The eurypterids, while appropriately small-bodied, are morphologically problematic. They appear more like generic segmented arthropods or oversized isopods than proper adelophthalmid eurypterids. Adelophthalmus had a distinctive prosoma (head shield), swimming paddles on the sixth appendage pair, and a characteristic telson shape. The creatures depicted lack visible swimming appendages and have body plans that are too uniformly segmented and elongate, with legs that look more like walking limbs of a terrestrial arthropod. The glossy, heavily segmented appearance reads more like trilobites or large aquatic isopods than eurypterids.
The bivalves scattered across the mudflat are overly large, rounded, and conspicuous—they look more like modern freshwater mussels than Carboniferous Edmondia, which were typically more elongate-oval in outline. The sheer number of intact, symmetrically placed shells also feels somewhat artificial. The small white objects scattered around could represent ostracods or shell fragments, which is a nice touch if intentional. The bright yellow object near the right side of the frame is unexplained and potentially anachronistic or simply an artifact.
The vegetation is the image's strongest element but still could be improved. The tall, relatively bare trunks in the background are reasonable for lycopsids like Lepidodendron or Sigillaria, though they lack the characteristic diamond or hexagonal leaf-scar patterning that would make them unmistakably Carboniferous. The large fern-like fronds on the right could represent tree ferns or seed ferns, which is appropriate. Some of the foliage, however, has a slightly too modern tropical appearance—more like cycads or contemporary tree ferns than distinctly Paleozoic forms. Calamites are not clearly distinguishable in the scene.
I largely agree with GPT's assessment. Their observation about the eurypterids being too 'shrimp-like' is on point, though I would go further—they're not just generically arthropod-like, they fundamentally lack key eurypterid morphological features (the broad prosoma, swimming paddles, and proper telson). GPT's caption approval is well-justified; the caption is scientifically sound, accurately describing Adelophthalmus ecology, the brackish setting, and the appropriate flora. The taxonomic and temporal details are all correct. The caption is doing more heavy lifting than the image in terms of scientific accuracy.
The bivalves scattered across the mudflat are overly large, rounded, and conspicuous—they look more like modern freshwater mussels than Carboniferous Edmondia, which were typically more elongate-oval in outline. The sheer number of intact, symmetrically placed shells also feels somewhat artificial. The small white objects scattered around could represent ostracods or shell fragments, which is a nice touch if intentional. The bright yellow object near the right side of the frame is unexplained and potentially anachronistic or simply an artifact.
The vegetation is the image's strongest element but still could be improved. The tall, relatively bare trunks in the background are reasonable for lycopsids like Lepidodendron or Sigillaria, though they lack the characteristic diamond or hexagonal leaf-scar patterning that would make them unmistakably Carboniferous. The large fern-like fronds on the right could represent tree ferns or seed ferns, which is appropriate. Some of the foliage, however, has a slightly too modern tropical appearance—more like cycads or contemporary tree ferns than distinctly Paleozoic forms. Calamites are not clearly distinguishable in the scene.
I largely agree with GPT's assessment. Their observation about the eurypterids being too 'shrimp-like' is on point, though I would go further—they're not just generically arthropod-like, they fundamentally lack key eurypterid morphological features (the broad prosoma, swimming paddles, and proper telson). GPT's caption approval is well-justified; the caption is scientifically sound, accurately describing Adelophthalmus ecology, the brackish setting, and the appropriate flora. The taxonomic and temporal details are all correct. The caption is doing more heavy lifting than the image in terms of scientific accuracy.
Grok
Image:
Regenerate
Caption:
Approved
Mar 27, 2026
The image captures a plausible overall setting for a Late Carboniferous estuarine lagoon with organic-rich mudflats, shallow brackish water, scattered bivalve shells, woody debris, and a misty tropical swamp forest backdrop, evoking the coal-forming wetlands of equatorial Euramerica. Dark sediments and peat-like tones are geologically accurate, and the quiet shoreline habitat aligns well with the period's coastal ecosystems. However, the central subjects—supposedly small Adelophthalmus eurypterids—are significantly misrepresented. These creatures appear as elongated, multi-legged (dozens of appendages) arthropods resembling giant isopods, centipedes, or even millipedes like Arthropleura, rather than the distinctive eurypterid morphology: broad crescentic prosoma, chelicerae, pedipalps, six pairs of walking appendages with the last as flattened paddles, segmented opisthosoma, and a spiny telson. No swimming paddles, head shields, or scorpion-like features are evident, making them anachronistic and biologically implausible for Adelophthalmus, one of the last euryhaline eurypterids. The bivalves resemble generic small clams (possibly fitting Edmondia), and tiny white specks could pass for ostracods, but the yellow object on the right mudflat is an unexplained artifact. Vegetation includes tall scalelike trunks suggestive of Calamites and lycopsids, plus frondose ferns/seed ferns, but lacks diagnostic Carboniferous traits like diamond-patterned leaf scars on Lepidodendron/Sigillaria, jointed Calamites nodes, or marattialean tree fern architecture—appearing too generically modern tropical (palm-like crowns). Visual coherence is good, but scientific inaccuracies in fauna and flora specificity demand a complete regenerate for accuracy.
The caption is factually precise and well-contextualized: Late Carboniferous (310–300 Ma) timing matches Westphalian stage; Adelophthalmus was indeed among the final, adaptable eurypterids in brackish habitats; Edmondia is a correct non-marine/estuarine bivalve; ostracod mats fit benthic assemblages; flora (lycopsids, Calamites, tree ferns, seed ferns) dominates Euramerican coal swamps; and the equatorial wetland interface is spot-on for peat accumulation and tropical biodiversity peaks. No misleading claims, appropriate detail level, fully consistent with paleontology.
I largely agree with GPT and Claude's image 'adjust' votes and caption approval, as they correctly flagged eurypterid morphology (shrimp/isopod-like, lacking prosoma/paddles/telson) and generic vegetation lacking Carboniferous hallmarks (e.g., leaf scars, Calamites structure). However, I diverge to 'regenerate' for the image due to the severity of faunal misrepresentation—the eurypterids dominate compositionally yet are unrecognizable, evoking wrong taxa (myriapods over merostomes), which minor prompt tweaks may not fully salvage without reemphasizing anatomy from references. They missed highlighting the excessive leg count (far beyond 12 appendages) mimicking terrestrial myriapods, incongruent for aquatic eurypterids, and the yellow anomaly as a potential artifact.
The caption is factually precise and well-contextualized: Late Carboniferous (310–300 Ma) timing matches Westphalian stage; Adelophthalmus was indeed among the final, adaptable eurypterids in brackish habitats; Edmondia is a correct non-marine/estuarine bivalve; ostracod mats fit benthic assemblages; flora (lycopsids, Calamites, tree ferns, seed ferns) dominates Euramerican coal swamps; and the equatorial wetland interface is spot-on for peat accumulation and tropical biodiversity peaks. No misleading claims, appropriate detail level, fully consistent with paleontology.
I largely agree with GPT and Claude's image 'adjust' votes and caption approval, as they correctly flagged eurypterid morphology (shrimp/isopod-like, lacking prosoma/paddles/telson) and generic vegetation lacking Carboniferous hallmarks (e.g., leaf scars, Calamites structure). However, I diverge to 'regenerate' for the image due to the severity of faunal misrepresentation—the eurypterids dominate compositionally yet are unrecognizable, evoking wrong taxa (myriapods over merostomes), which minor prompt tweaks may not fully salvage without reemphasizing anatomy from references. They missed highlighting the excessive leg count (far beyond 12 appendages) mimicking terrestrial myriapods, incongruent for aquatic eurypterids, and the yellow anomaly as a potential artifact.
Gemini
Image:
Adjust
Caption:
Approved
Mar 27, 2026
The image successfully captures the atmospheric 'vibe' of a Late Carboniferous swamp—misty, dark, and organic—but fails significantly on the anatomical and botanical specifics required for a scientific reconstruction. The primary issue is the depiction of Adelophthalmus. As my colleagues noted, these creatures lack the diagnostic 'swimming paddles' (the 6th pair of appendages) and the distinct prosoma characteristic of eurypterids. I agree with Grok that they appear more like myriapods or isopods due to an excessive number of uniform walking legs, which is a major morphological error. Additionally, the flora is too 'generic tropical'; without the diamond-patterned leaf scars of Lepidodendron or the jointed nodes of Calamites, the image loses its Carboniferous identity and looks more like a modern mangrove or cypress swamp.
I approve the caption without reservation. It is scientifically rigorous, correctly identifying the temporal range, the specific brackish ecology, and the relevant taxa (Adelophthalmus, Edmondia, and the dominant flora). The text provides the necessary context that the image currently lacks.
In response to the other committee members: I agree with Claude and Grok regarding the 'yellow artifact' and the morphological failures of the fauna. I find GPT’s assessment of the eurypterids as 'shrimp-like' to be a bit of an understatement—they are fundamentally the wrong class of arthropod in their current rendering. I support an 'adjust' vote rather than 'regenerate' because the lighting, sediment texture, and composition are excellent foundations; however, the prompt must be significantly tightened to demand specific eurypterid anatomy (paddles, head shield) and lycopsid bark patterns to move from a 'generic swamp' to a 'historically accurate Carboniferous lagoon.'
I approve the caption without reservation. It is scientifically rigorous, correctly identifying the temporal range, the specific brackish ecology, and the relevant taxa (Adelophthalmus, Edmondia, and the dominant flora). The text provides the necessary context that the image currently lacks.
In response to the other committee members: I agree with Claude and Grok regarding the 'yellow artifact' and the morphological failures of the fauna. I find GPT’s assessment of the eurypterids as 'shrimp-like' to be a bit of an understatement—they are fundamentally the wrong class of arthropod in their current rendering. I support an 'adjust' vote rather than 'regenerate' because the lighting, sediment texture, and composition are excellent foundations; however, the prompt must be significantly tightened to demand specific eurypterid anatomy (paddles, head shield) and lycopsid bark patterns to move from a 'generic swamp' to a 'historically accurate Carboniferous lagoon.'
Matania
Summary
Image:
Adjust
Caption:
Approved
The committee agrees that the image gets the broad setting right: it convincingly depicts a quiet Late Carboniferous estuarine or lagoonal shoreline with dark organic-rich mud, shallow brackish water, peat-like tones, scattered shell material, woody debris, and a swamp forest backdrop appropriate to equatorial Euramerican coal-forming wetlands. The small body size and shoreline placement of the supposed Adelophthalmus are also broadly sensible in ecological terms. The caption is unanimously regarded as scientifically strong, accurate in age, setting, ecology, and taxonomic content.
For the IMAGE, the committee identified the following issues: (1) the Adelophthalmus are morphologically incorrect overall and are not clearly recognizable as adelophthalmid eurypterids; (2) they look too uniformly shrimp-like, generic arthropod-like, or glossy; (3) they resemble oversized isopods, trilobites, centipedes, millipedes, myriapods, or Arthropleura more than eurypterids; (4) their body plans are too uniformly segmented and too elongate; (5) they appear to have an excessive number of appendages/legs, far beyond the expected eurypterid condition; (6) the limbs read as generalized walking legs or even terrestrial arthropod legs rather than eurypterid appendages; (7) the distinctive broad/crescentic prosoma or head shield is missing or inadequately expressed; (8) the characteristic swimming paddles on the sixth appendage pair are missing, unclear, or not diagnostic; (9) telson shape/details are weak, incorrect, or not diagnostic; (10) scorpion-like/merostome features are insufficiently evident; (11) the animals are compositionally central enough that this anatomical failure materially weakens scientific accuracy. (12) The shell assemblage is too dominated by large, conspicuous, intact whole bivalves for a quiet muddy coal-swamp margin; (13) the bivalves are overly large, rounded, and conspicuous; (14) they look too much like modern freshwater mussels or generic clams rather than Edmondia; (15) Edmondia should be more elongate-oval in outline; (16) the abundance and symmetrical placement of intact shells feels artificial; (17) more subtle fragmentary shell debris and microbial textures would be preferable. (18) The small white specks may work as ostracods or shell debris, but their identity should be made clearer if intended. (19) There is a bright yellow object on the right mudflat that is unexplained and likely an artifact/anachronistic visual intrusion and should be removed. (20) Vegetation is broadly appropriate but too generic or too modern-tropical in appearance; (21) some trunks and fronds edge toward a generic tropical swamp, mangrove, cypress swamp, cycad grove, or modern tree-fern forest rather than a distinctly Carboniferous flora; (22) lycopsid trunks lack clear Lepidodendron/Sigillaria-style diamond or hexagonal leaf-scar patterning; (23) Calamites are not clearly distinguishable and/or lack obvious jointed nodes and architecture; (24) some palm-like crowns or modern-looking fern architecture reduce Paleozoic specificity; (25) marattialean/tree-fern and seed-fern forms should be more clearly Paleozoic rather than modern ornamental fern-like. Overall, the image has a good environmental foundation but needs anatomical and botanical tightening to become a reliable scientific reconstruction.
For the CAPTION, the committee found no factual errors, anachronisms, inconsistencies, or misleading claims. Specifically, no reviewer identified any problem with the stated Late Carboniferous age, the estuarine lagoon interpretation, the ecological characterization of Adelophthalmus as among the last and adaptable eurypterids, the mention of Edmondia bivalves and ostracod-rich microbial mats, or the listed wetland flora of lycopsids, Calamites, tree ferns, and seed ferns in equatorial Euramerica.
Verdict: adjust the image, approve the caption. The majority image vote was adjust, and although one reviewer argued for regeneration because the eurypterids are currently so inaccurate, the scene's lighting, sedimentology, composition, and overall paleoenvironment are strong enough to preserve. However, the fauna and flora must be corrected decisively: the central animals need to be unmistakably adelophthalmid eurypterids, the shell bed must be toned down and made taxonomically/tafonomically more plausible, the yellow artifact must be removed, and the vegetation must gain clearer Carboniferous hallmarks. The caption already meets committee standards and requires no change.
For the IMAGE, the committee identified the following issues: (1) the Adelophthalmus are morphologically incorrect overall and are not clearly recognizable as adelophthalmid eurypterids; (2) they look too uniformly shrimp-like, generic arthropod-like, or glossy; (3) they resemble oversized isopods, trilobites, centipedes, millipedes, myriapods, or Arthropleura more than eurypterids; (4) their body plans are too uniformly segmented and too elongate; (5) they appear to have an excessive number of appendages/legs, far beyond the expected eurypterid condition; (6) the limbs read as generalized walking legs or even terrestrial arthropod legs rather than eurypterid appendages; (7) the distinctive broad/crescentic prosoma or head shield is missing or inadequately expressed; (8) the characteristic swimming paddles on the sixth appendage pair are missing, unclear, or not diagnostic; (9) telson shape/details are weak, incorrect, or not diagnostic; (10) scorpion-like/merostome features are insufficiently evident; (11) the animals are compositionally central enough that this anatomical failure materially weakens scientific accuracy. (12) The shell assemblage is too dominated by large, conspicuous, intact whole bivalves for a quiet muddy coal-swamp margin; (13) the bivalves are overly large, rounded, and conspicuous; (14) they look too much like modern freshwater mussels or generic clams rather than Edmondia; (15) Edmondia should be more elongate-oval in outline; (16) the abundance and symmetrical placement of intact shells feels artificial; (17) more subtle fragmentary shell debris and microbial textures would be preferable. (18) The small white specks may work as ostracods or shell debris, but their identity should be made clearer if intended. (19) There is a bright yellow object on the right mudflat that is unexplained and likely an artifact/anachronistic visual intrusion and should be removed. (20) Vegetation is broadly appropriate but too generic or too modern-tropical in appearance; (21) some trunks and fronds edge toward a generic tropical swamp, mangrove, cypress swamp, cycad grove, or modern tree-fern forest rather than a distinctly Carboniferous flora; (22) lycopsid trunks lack clear Lepidodendron/Sigillaria-style diamond or hexagonal leaf-scar patterning; (23) Calamites are not clearly distinguishable and/or lack obvious jointed nodes and architecture; (24) some palm-like crowns or modern-looking fern architecture reduce Paleozoic specificity; (25) marattialean/tree-fern and seed-fern forms should be more clearly Paleozoic rather than modern ornamental fern-like. Overall, the image has a good environmental foundation but needs anatomical and botanical tightening to become a reliable scientific reconstruction.
For the CAPTION, the committee found no factual errors, anachronisms, inconsistencies, or misleading claims. Specifically, no reviewer identified any problem with the stated Late Carboniferous age, the estuarine lagoon interpretation, the ecological characterization of Adelophthalmus as among the last and adaptable eurypterids, the mention of Edmondia bivalves and ostracod-rich microbial mats, or the listed wetland flora of lycopsids, Calamites, tree ferns, and seed ferns in equatorial Euramerica.
Verdict: adjust the image, approve the caption. The majority image vote was adjust, and although one reviewer argued for regeneration because the eurypterids are currently so inaccurate, the scene's lighting, sedimentology, composition, and overall paleoenvironment are strong enough to preserve. However, the fauna and flora must be corrected decisively: the central animals need to be unmistakably adelophthalmid eurypterids, the shell bed must be toned down and made taxonomically/tafonomically more plausible, the yellow artifact must be removed, and the vegetation must gain clearer Carboniferous hallmarks. The caption already meets committee standards and requires no change.
Other languages
- Français: Scorpions de mer Adelophthalmus dans une lagune saumâtre Carbonifère
- Español: Escorpiones marinos Adelophthalmus en una laguna salobre del Carbonífero
- Português: Escorpiões marinhos Adelophthalmus em uma lagoa salobra do Carbonífero
- Deutsch: Kleine Adelophthalmus-See-Skorpione in einer brackigen Lagune des Karbons
- العربية: عقارب البحر أديلوفثالموس الصغيرة في بحيرة مالحة من العصر الفحمي
- हिन्दी: कार्बोनिफेरस खारी लैगून में छोटे एडेलोफ्थाल्मस समुद्री बिच्छू
- 日本語: 石炭紀の汽水ラグーンに生息する小型のウミサソリ
- 한국어: 석탄기 기수호에 서식하는 작은 바다전갈 아델로프탈무스
- Italiano: Scorpioni di mare Adelophthalmus in una laguna salmastra Carbonifera
- Nederlands: Kleine Adelophthalmus zeeschorpioenen in een brakke lagune uit het Carboon
That said, the image would benefit from refinement. The eurypterids look somewhat too uniformly shrimp-like and glossy, with legs and body proportions that read more like generalized arthropods than well-observed Adelophthalmus; their paddles and telson details are not especially diagnostic. The shell bed is also a little too dominated by large conspicuous whole bivalves for a quiet muddy coal-swamp margin, where more fragmentary shell accumulations and microbial textures would likely be subtler. In the vegetation, the forest is broadly correct, but some trunks and fronds feel slightly too modern in growth form, edging toward a generic tropical swamp rather than distinctly Carboniferous architecture; emphasizing more obvious Lepidodendron/Sigillaria patterning and Calamites structure would improve period specificity.
The caption is strong and largely accurate. Adelophthalmus is indeed among the last surviving and ecologically flexible eurypterids, and a brackish estuarine lagoon at the edge of coal-swamp forest in equatorial Euramerica around 310–300 Ma is a good reconstruction. Mentioning Edmondia bivalves, ostracods, microbial mats, lycopsids, Calamites, tree ferns, and seed ferns is consistent with the setting and time interval, and the description appropriately links coastal waters with adjacent peat-forming wetlands.
I would therefore approve the caption as written, while recommending minor image adjustments to make the animals more anatomically recognizable as adelophthalmid eurypterids and the flora more unmistakably Carboniferous rather than generically swamp-tropical. No major anachronisms are present, but the visual specificity could be improved.