In a dim, waterlogged coal-swamp of late Carboniferous Euramerica, about 310–305 million years ago, a giant Arthropleura—an enormous millipede relative reaching roughly 2 meters in length—moves across rotting Lepidodendron bark and fern litter as small roachoid insects scatter from its path. Towering lycopsids such as Lepidodendron and Sigillaria rise above the forest floor, joined by tree ferns including Psaronius and horsetail-like Calamites in the saturated mire. This scene captures the humid equatorial peat forests that later formed many of Earth’s coal deposits, an ecosystem famous for its giant arthropods and dense spore-bearing vegetation long before flowering plants evolved.
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 captures the general mood of a Carboniferous coal swamp effectively: the dim, humid atmosphere, standing water, fallen logs, and a large Arthropleura are all appropriate. The Arthropleura itself is reasonably well-rendered with its segmented exoskeleton, though it appears somewhat dark and shiny — more beetle-like than the flattened, broad-bodied form typically reconstructed for this animal. Its size relative to the surrounding debris seems plausible for a ~2 meter specimen.
However, I share and amplify the GPT reviewer's concerns about the vegetation. The most glaring issue is the scattered leaf litter on the forest floor: these are unmistakably broad, angiosperm-style leaves (they look like dried maple or deciduous hardwood leaves), which is a significant anachronism — flowering plants did not exist for another ~180 million years. The Carboniferous forest floor should be littered with bark fragments, narrow frond pinnules, lycopsid leaf cushion debris, and decaying plant matter, not broad dicot-like leaves. The green plants in the mid-ground that look distinctly like bamboo (segmented green stems with leaf tufts) are also problematic; while Calamites were sphenopsids with jointed stems, bamboo is a grass (angiosperm) and the visual impression here is too modern. Some of the tree ferns in the background are reasonably convincing as Psaronius-like forms, and some of the taller trunks with diamond/scale patterns could pass for lycopsids, but several trees look more like modern tropical palms than any Carboniferous taxon. The overall canopy structure reads somewhat too modern-tropical.
The caption is scientifically accurate and well-written. The date range of 310–305 Ma for late Carboniferous is appropriate. The taxa mentioned — Arthropleura, Lepidodendron, Sigillaria, Psaronius, Calamites — are all correct for this setting. The ecological context about coal formation and the pre-angiosperm nature of the flora is accurate. The mention of roachoid insects is paleontologically sound, as blattodean relatives were indeed common in Carboniferous ecosystems.
I agree with the GPT reviewer's assessment almost entirely. If anything, I would emphasize even more strongly that the broad deciduous-looking leaf litter is the single most jarring anachronism in the image and should be the top priority for correction. The bamboo-like Calamites and some palm-like tree forms are secondary concerns. These issues collectively warrant adjustment rather than full regeneration, as the overall composition and atmosphere are strong.
However, I share and amplify the GPT reviewer's concerns about the vegetation. The most glaring issue is the scattered leaf litter on the forest floor: these are unmistakably broad, angiosperm-style leaves (they look like dried maple or deciduous hardwood leaves), which is a significant anachronism — flowering plants did not exist for another ~180 million years. The Carboniferous forest floor should be littered with bark fragments, narrow frond pinnules, lycopsid leaf cushion debris, and decaying plant matter, not broad dicot-like leaves. The green plants in the mid-ground that look distinctly like bamboo (segmented green stems with leaf tufts) are also problematic; while Calamites were sphenopsids with jointed stems, bamboo is a grass (angiosperm) and the visual impression here is too modern. Some of the tree ferns in the background are reasonably convincing as Psaronius-like forms, and some of the taller trunks with diamond/scale patterns could pass for lycopsids, but several trees look more like modern tropical palms than any Carboniferous taxon. The overall canopy structure reads somewhat too modern-tropical.
The caption is scientifically accurate and well-written. The date range of 310–305 Ma for late Carboniferous is appropriate. The taxa mentioned — Arthropleura, Lepidodendron, Sigillaria, Psaronius, Calamites — are all correct for this setting. The ecological context about coal formation and the pre-angiosperm nature of the flora is accurate. The mention of roachoid insects is paleontologically sound, as blattodean relatives were indeed common in Carboniferous ecosystems.
I agree with the GPT reviewer's assessment almost entirely. If anything, I would emphasize even more strongly that the broad deciduous-looking leaf litter is the single most jarring anachronism in the image and should be the top priority for correction. The bamboo-like Calamites and some palm-like tree forms are secondary concerns. These issues collectively warrant adjustment rather than full regeneration, as the overall composition and atmosphere are strong.
Grok
Image:
Adjust
Caption:
Approved
Mar 27, 2026
The image effectively captures the dim, humid, waterlogged atmosphere of a late Carboniferous coal swamp, with standing water, fallen logs, muddy terrain, and a plausibly scaled Arthropleura (roughly 2m long, segmented exoskeleton) crawling through debris—core elements that align well with equatorial Euramerican forests. The giant arthropod's form is reasonably accurate, resembling reconstructions as a flattened, multi-legged myriapod relative, and small fern-like plants in the foreground add to the spore-dominated flora vibe. Background trunks show some diamond-patterned bark suggestive of lycopsids like Lepidodendron, and reed-like stems evoke Calamites, while the misty, low-light canopy contributes to visual coherence and plausibility for a peat-forming mire.
However, several botanical anachronisms warrant adjustment rather than regeneration. The most prominent issue is the scattered broad, orange-brown leaves on the forest floor, which strongly resemble modern angiosperm foliage (e.g., maple or oak leaves) rather than the expected Carboniferous litter of narrow lycopsid leaf scars, frond pinnules, bark scales, and decaying spores—flowering plants were absent until the Cretaceous. Some midground stems appear bamboo-like (grassy angiosperms, post-Carboniferous), and several background trees have palm- or cycad-like fronds that feel too modern-tropical compared to the whisk-like crowns of Psaronius or cone-bearing lycopsids. These are refinable via prompt tweaks emphasizing taxon-specific morphologies (e.g., 'no broad leaves, only scale bark and narrow fronds'). No major geological or faunal errors, and the scene avoids human/modern intrusions.
The caption is factually precise and consistent: correct taxa (Arthropleura, Lepidodendron, Sigillaria, Psaronius, Calamites), timeframe (310–305 Ma, late Carboniferous), location (Euramerica coal swamps), ecology (peat forests forming coal, giant arthropods, pre-angiosperm spores), and details like roachoids and rotting bark. It provides appropriate context without misleading claims. I fully agree with GPT and Claude's assessments—the angiosperm-like leaf litter is the standout flaw they highlighted, correctly flagged as jarring, and the bamboo/palm impressions are secondary but valid. They missed nothing significant; their analysis aligns with mine, justifying image adjustment while approving the strong caption.
However, several botanical anachronisms warrant adjustment rather than regeneration. The most prominent issue is the scattered broad, orange-brown leaves on the forest floor, which strongly resemble modern angiosperm foliage (e.g., maple or oak leaves) rather than the expected Carboniferous litter of narrow lycopsid leaf scars, frond pinnules, bark scales, and decaying spores—flowering plants were absent until the Cretaceous. Some midground stems appear bamboo-like (grassy angiosperms, post-Carboniferous), and several background trees have palm- or cycad-like fronds that feel too modern-tropical compared to the whisk-like crowns of Psaronius or cone-bearing lycopsids. These are refinable via prompt tweaks emphasizing taxon-specific morphologies (e.g., 'no broad leaves, only scale bark and narrow fronds'). No major geological or faunal errors, and the scene avoids human/modern intrusions.
The caption is factually precise and consistent: correct taxa (Arthropleura, Lepidodendron, Sigillaria, Psaronius, Calamites), timeframe (310–305 Ma, late Carboniferous), location (Euramerica coal swamps), ecology (peat forests forming coal, giant arthropods, pre-angiosperm spores), and details like roachoids and rotting bark. It provides appropriate context without misleading claims. I fully agree with GPT and Claude's assessments—the angiosperm-like leaf litter is the standout flaw they highlighted, correctly flagged as jarring, and the bamboo/palm impressions are secondary but valid. They missed nothing significant; their analysis aligns with mine, justifying image adjustment while approving the strong caption.
Gemini
Image:
Adjust
Caption:
Approved
Mar 27, 2026
The image effectively captures the atmospheric 'feel' of a Carboniferous coal swamp—the dim lighting, stagnant water, and the impressive scale of Arthropleura are all well-executed. The Arthropleura’s placement and size relative to the environment are scientifically plausible for the Euramerican tropics. However, I agree with the previous reviewers that the vegetation contains significant anachronisms. The most egregious error is the leaf litter; the presence of broad, flat, deciduous-style leaves is a major botanical inaccuracy, as these belong to angiosperms (flowering plants) which did not exist in the Carboniferous. The debris should instead consist of diamond-patterned bark fragments, needle-like lycopsid leaves, and fern fronds.
Additionally, the 'bamboo-like' appearance of the Calamites and the 'palm-like' silhouettes in the background are too modern. While Calamites were indeed jointed, their morphology is distinct from modern grasses, and palms would not evolve for hundreds of millions of years. The AI has clearly defaulted to a modern 'tropical jungle' template rather than a strictly spore-bearing Carboniferous flora. The caption, however, is excellent. It correctly identifies the timeframe (310–305 Ma), the specific taxa (Lepidodendron, Sigillaria, Psaronius, Calamites), and the ecological significance of the peat forests. I concur with my colleagues: the caption is ready for approval, but the image requires adjustment to remove broad-leaf litter and refine the plant morphologies to be historically accurate.
Additionally, the 'bamboo-like' appearance of the Calamites and the 'palm-like' silhouettes in the background are too modern. While Calamites were indeed jointed, their morphology is distinct from modern grasses, and palms would not evolve for hundreds of millions of years. The AI has clearly defaulted to a modern 'tropical jungle' template rather than a strictly spore-bearing Carboniferous flora. The caption, however, is excellent. It correctly identifies the timeframe (310–305 Ma), the specific taxa (Lepidodendron, Sigillaria, Psaronius, Calamites), and the ecological significance of the peat forests. I concur with my colleagues: the caption is ready for approval, but the image requires adjustment to remove broad-leaf litter and refine the plant morphologies to be historically accurate.
Matania
Summary
Image:
Adjust
Caption:
Approved
The committee agrees that the scene successfully captures the overall setting of a late Carboniferous equatorial coal swamp: a dim, humid, waterlogged mire with standing water, mud, fallen logs, abundant spore-bearing vegetation, and a plausibly large Arthropleura in an appropriate forest-floor habitat. Reviewers also agree that the image contains no glaring faunal or cultural anachronisms such as mammals, humans, buildings, or flowering shrubs/flowers, and that the Arthropleura is broadly plausible in scale and general placement. The caption is unanimously regarded as scientifically strong, with accurate time range, setting, taxa, and ecological context.
For the IMAGE, the complete list of issues identified by the committee is: 1. The most serious problem is the forest-floor litter: numerous scattered broad, flat, orange-brown leaves resemble modern angiosperm/dicot deciduous leaves (described by reviewers as maple-, oak-, or hardwood-like), which are a major anachronism because flowering plants did not exist in the Carboniferous. 2. The ground debris therefore looks too modern and should instead consist of Carboniferous-appropriate material such as rotting lycopsid bark, narrow fern/tree-fern frond fragments and pinnules, sphenopsid debris, lycopsid leaf-cushion fragments, woody/bark scales, and amorphous peat/muck. 3. Several upright segmented green stems read visually as bamboo; although likely intended as Calamites, their current design creates a distinctly bamboo/grass-like angiosperm impression and needs correction. 4. Some vegetation has a generalized modern tropical-jungle look rather than clearly taxon-specific Carboniferous morphology. 5. Several background trees or crowns look too palm-like, which is anachronistic for the Carboniferous. 6. Some plants resemble modern tree ferns or palms rather than clearly Carboniferous forms such as Psaronius, Lepidodendron, Sigillaria, and Calamites. 7. The canopy structure overall reads somewhat too modern-tropical rather than a distinct Carboniferous coal-swamp flora. 8. The Arthropleura is broadly acceptable, but one reviewer noted it appears somewhat too dark and shiny, giving it a slightly beetle-like look instead of the flatter, broader-bodied appearance usually reconstructed for Arthropleura. 9. One reviewer also noted that the visual is less taxonomically precise than the caption, meaning the image should better distinguish lycopsids, sphenopsids, and tree ferns from modern analogues.
For the CAPTION, the complete list of issues identified by the committee is: none requiring correction. Reviewers consistently found it scientifically accurate and well written. Minor commentary was only that the caption is more taxonomically precise than the current image, not that the caption itself contains an error. Specific praised elements include: Arthropleura reaching roughly 2 meters, late Carboniferous age of about 310–305 Ma, Euramerican coal-swamp context, the listed taxa Lepidodendron, Sigillaria, Psaronius, and Calamites, the reference to roachoid insects, the note that these peat forests later formed major coal deposits, and the statement that this ecosystem predates flowering plants.
Final verdict: adjust the image, approve the caption. The image succeeds compositionally and atmospherically and does not need full regeneration, but it contains repeated botanical anachronisms and overly modernized plant designs that undermine period fidelity—especially the broad angiosperm-like leaf litter and bamboo-/palm-like vegetation. These are localized, correctable design problems within an otherwise strong Carboniferous swamp scene. The caption needs no changes.
For the IMAGE, the complete list of issues identified by the committee is: 1. The most serious problem is the forest-floor litter: numerous scattered broad, flat, orange-brown leaves resemble modern angiosperm/dicot deciduous leaves (described by reviewers as maple-, oak-, or hardwood-like), which are a major anachronism because flowering plants did not exist in the Carboniferous. 2. The ground debris therefore looks too modern and should instead consist of Carboniferous-appropriate material such as rotting lycopsid bark, narrow fern/tree-fern frond fragments and pinnules, sphenopsid debris, lycopsid leaf-cushion fragments, woody/bark scales, and amorphous peat/muck. 3. Several upright segmented green stems read visually as bamboo; although likely intended as Calamites, their current design creates a distinctly bamboo/grass-like angiosperm impression and needs correction. 4. Some vegetation has a generalized modern tropical-jungle look rather than clearly taxon-specific Carboniferous morphology. 5. Several background trees or crowns look too palm-like, which is anachronistic for the Carboniferous. 6. Some plants resemble modern tree ferns or palms rather than clearly Carboniferous forms such as Psaronius, Lepidodendron, Sigillaria, and Calamites. 7. The canopy structure overall reads somewhat too modern-tropical rather than a distinct Carboniferous coal-swamp flora. 8. The Arthropleura is broadly acceptable, but one reviewer noted it appears somewhat too dark and shiny, giving it a slightly beetle-like look instead of the flatter, broader-bodied appearance usually reconstructed for Arthropleura. 9. One reviewer also noted that the visual is less taxonomically precise than the caption, meaning the image should better distinguish lycopsids, sphenopsids, and tree ferns from modern analogues.
For the CAPTION, the complete list of issues identified by the committee is: none requiring correction. Reviewers consistently found it scientifically accurate and well written. Minor commentary was only that the caption is more taxonomically precise than the current image, not that the caption itself contains an error. Specific praised elements include: Arthropleura reaching roughly 2 meters, late Carboniferous age of about 310–305 Ma, Euramerican coal-swamp context, the listed taxa Lepidodendron, Sigillaria, Psaronius, and Calamites, the reference to roachoid insects, the note that these peat forests later formed major coal deposits, and the statement that this ecosystem predates flowering plants.
Final verdict: adjust the image, approve the caption. The image succeeds compositionally and atmospherically and does not need full regeneration, but it contains repeated botanical anachronisms and overly modernized plant designs that undermine period fidelity—especially the broad angiosperm-like leaf litter and bamboo-/palm-like vegetation. These are localized, correctable design problems within an otherwise strong Carboniferous swamp scene. The caption needs no changes.
Other languages
- Français: Mille-pattes géant Arthropleura sur le sol du Carbonifère
- Español: Milpiés gigante Arthropleura en el suelo del Carbonífero
- Português: Milpés gigante Arthropleura no solo florestal do Carbonífero
- Deutsch: Riesiger Tausendfüßler Arthropleura auf dem Waldboden des Karbons
- العربية: ألفية الأرجل العملاقة أرثروبليورا على أرض غابة الكربوني
- हिन्दी: कार्बोनिफेरस वन तल पर विशाल आर्थ्रोप््लुइरा सहस्रपाद
- 日本語: 石炭紀の林床を這う巨大な多足類アースロプレウラ
- 한국어: 석탄기 숲 바닥을 지나가는 거대한 아르트로플레우라
- Italiano: Millepiedi gigante Arthropleura sul suolo forestale del Carbonifero
- Nederlands: Reusachtige Arthropleura miljoenpoot op de bodem van het Carboon-bos
That said, some botanical details look a bit too modernized or generalized and would benefit from refinement. Several plants resemble later or modern tree ferns, palms, or even bamboo-like canes rather than clearly Carboniferous taxa; the upright segmented stems likely intend to represent Calamites, but they read visually like bamboo, which is an anachronistic impression. The broad scattered leaf litter on the ground also looks more angiosperm-like than the typical shed material expected in a Carboniferous swamp, where bark, narrow fronds, and lycopsid debris would be more appropriate. So the image does not require a full redo, but some vegetation design choices should be adjusted to better distinguish lycopsids, sphenopsids, and tree ferns from modern tropical analogues.
The caption is scientifically solid. Arthropleura reaching roughly 2 meters, association with late Carboniferous Euramerican coal swamps, and the dominance of Lepidodendron, Sigillaria, Psaronius, and Calamites are all appropriate. The note that these peat forests later formed many coal deposits is accurate, and the statement that this predates flowering plants is also correct. Mention of small roachoid insects is plausible for the setting, even if they are not obvious in the image.
If anything, the caption is slightly more taxonomically precise than the visual itself, which is why the caption can be approved while the image merits adjustment. A revised image prompt should emphasize unmistakable Carboniferous plant morphology and avoid modern-looking leaf litter and bamboo-like forms.