A Federal Reserve Bank of Philadelphia working paper published this month found that public notifications of large crypto transfers were followed by sharply different trading activity on Bitcoin and Ethereum. Non-whale Bitcoin wallets traded in the alerted whale's direction most strongly during the first 15 minutes, while Ethereum participation remained comparatively stable. The study matched Whale Alert notification times with on-chain Bitcoin, Ethereum and Wrapped Bitcoin (WBTC) transfers through the end of 2025. It defined a whale wallet as one making at least one transfer worth more than $50 million, excluding large wallets linked to exchanges or smart contracts. After removing events with another whale transaction within two hours before or after, the sample contained 6,645 BTC and 5,075 ETH whale transactions. Following whale purchases, buy participation rose 14.81 percentage points among small wallets, 23.72 points among medium wallets and 3.50 points among large wallets. After whale sales, sell participation increased by 12.95, 29.52 and 2.95 points, respectively, before returning toward normal within an hour. Ethereum showed no comparable broad response; its clearest same-direction result involved the largest non-whale group after whale sales, while medium ETH sellers reached only the study's weaker 10% significance threshold. Whale alerts coincided with temporarily higher realized Bitcoin volatility, but by 24 hours the BTC volatility response to BTC and ETH alerts had reversed, while WBTC alerts had no statistically distinguishable effect. Ethereum volatility was lower after alerts. The authors attribute the difference to market structure, noting that Ethereum activity often passes through exchanges, smart contracts and layer-2 venues that aggregate user transactions into larger transfers. The pattern persisted after Ethereum's September 2022 shift to proof-of-stake (a blockchain consensus method). The evidence is observational: wallet-size categories are transaction-based proxies, one owner may control multiple addresses, exchanges were excluded, and the study identifies patterns around alerts rather than proving that alerts caused every response.