A block is a batch of transactions that a blockchain network records together, sealed with a header that points back to the block before it. Each block adds one page to the shared public ledger, and the chain of these pages, each referring to its parent, is what gives a blockchain its name. Blocks are produced by miners on proof-of-work networks and by validators on proof-of-stake networks.
A block has two parts: a small header that summarises it and a body that holds the transactions. Every node checks a new block against the network's rules before accepting it, which is how thousands of independent computers keep identical copies of the ledger without a central bookkeeper. The block is the basic unit of consensus; terms such as block height, block reward and stale block all describe a block's position, payment or fate.
A block in plain terms
Think of a long receipt book kept in a shop where nobody trusts anyone. Every ten minutes or so, someone tears off a filled page, staples it to the previous one and writes a short summary at the top that includes a fingerprint of the previous page. If anyone later tries to change an old page, its fingerprint changes, the summary on the next page no longer matches, and everyone can see the tampering.
A block is one of those stapled pages. The list of sales is the body. The summary at the top, with the fingerprint of the page before it, is the header. Because every page points to the one before, the whole book can be checked by anyone, page by page, back to the very first one.
What a block contains
On the Bitcoin blockchain the header is exactly 80 bytes and holds six fields:
- The version number. It signals which set of validation rules the block follows.
- The previous block hash. This 32-byte fingerprint of the parent block's header is what links the chain together.
- The Merkle root. A single hash that summarises every transaction in the body, built with a Merkle tree.
- The timestamp. It must be later than the median time of the previous 11 blocks, and nodes reject a block dated more than two hours ahead of their own clock.
- The difficulty target (nBits). A compact encoding of how low the header hash must be for the block to count.
- The nonce. A number miners change again and again in search of a valid hash.
The body starts with the coinbase transaction, which creates the block reward, followed by ordinary transfers that spend unspent outputs. The amount of data a block may carry is capped; the page on block size explains how Bitcoin's limit is measured in weight units since SegWit.
Blocks on the Ethereum blockchain look different. Their execution payload records a block number, a fee recipient, a state root, a receipts root, the gas limit and gas used, a base fee per gas and a transactions root, among other fields. Instead of a byte limit, an Ethereum block is limited by the total gas its transactions may consume.
How a block is made and accepted
Pending transactions wait in each node's mempool. A block producer picks transactions, usually those paying the highest fees, builds the body, computes the Merkle root and fills in the header.
On a proof-of-work chain the producer then hashes the header over and over, changing the nonce each time, until the result falls below the target set by mining difficulty. Bitcoin is tuned so that this happens about once every ten minutes on average. On Ethereum since the Merge of September 2022, time is divided into 12-second slots; one validator is chosen to propose a block in each slot, and other validators attest to it. A slot can be empty if the chosen validator is offline.
Every full node that receives the block re-checks it independently: the hash meets the target, every signature is valid, no coin is spent twice and the reward is not too large. Only then does the node add the block to its copy of the chain and pass it to its peers.
Why blocks are hard to change
Because each header contains the hash of its parent, altering a transaction in an old block changes that block's Merkle root, its header hash, and therefore the "previous block hash" stored in every later block. On a proof-of-work chain an attacker would have to redo the work for the altered block and every block after it, faster than the rest of the network; that is the scenario described as a 51% attack.
This is why recipients wait for confirmations: each new block on top of the one holding a payment makes reversal more expensive. Bitcoin's finality is therefore probabilistic, growing stronger with each block, while Ethereum's proof-of-stake design adds checkpoints that become final after roughly two epochs. Occasionally two valid blocks are found at nearly the same time; one becomes part of the main chain and the other becomes a stale block.
Blocks across networks
Block design is a trade-off between speed, capacity and the cost of running a node. Shorter intervals reduce waiting time but raise the share of competing blocks; larger blocks carry more transactions but demand more bandwidth and storage. Disagreements over this trade-off led to the 2017 split that created Bitcoin Cash, a hard fork with larger blocks.
Not every distributed ledger uses blocks at all. Some designs record transactions in a directed acyclic graph rather than a single chain, and many networks move activity onto a layer 2 system that settles in batches on the base chain. Individual blocks can be inspected with a block explorer, which shows each block's height, hash, size and transactions, back to the genesis block.