The Senate and the United States Constitution
Chapter 1: Two Senators per State
Chapter 2: Term Length
Chapter 3: Senate Classes
Chapter 4: Qualifications
Chapter 5: President of the Senate
Chapter 6: Other Senate Officers
Chapter 7: Senate Impeachment Trials
Chapter 8: Treaties
Chapter 9: Nominations
Chapter 10: Sources
The Senate of the United States shall be composed of two Senators from each State. [U.S. Constitution, Article I, section 3, clause 1]
During the summer of 1787, the delegates to the Constitutional Convention in Philadelphia established equal representation in the Senate and proportional representation in the House of Representatives. Called the “Great Compromise” or the “Connecticut Compromise,” the unique plan for congressional representation resolved the most controversial aspect of the drafting of the Constitution.
In the weeks before the Constitution’s framers agreed to the compromise, the delegates from the states with large populations argued that each state’s representation in the Senate should correspond to the size of the state. Large-state delegates promoted James Madison’s Virginia Plan, the document that was the basis for several of the clauses in the Constitution. Under this plan, the Senate and the House would base their membership on the same proportional “right of suffrage.” That is, the number of senators in each state would be determined by its population of free citizens and slaves. Large states, then, stood to gain the most seats in the Senate. As justification for this advantage, delegates noted that their states contributed more of the nation’s financial and defensive resources than small states, and therefore, required a greater say in government.
Small-state delegates hoped to protect states’ rights within a confederate system of government. Fearing the effects of majority rule, they demanded equal representation in Congress, as was practiced under the Articles of Confederation and assumed in William Paterson’s New Jersey Plan. In fact, some framers threatened to withdraw from the convention if a proportional representation measure passed.
Other delegates sought a compromise between large-state and small-state interests. As early as 1776, Connecticut’s Roger Sherman had suggested that Congress represent the people as well as the states. During the 1787 convention, Sherman proposed that House representation be based on the population, while in the Senate, the states would be equally represented. Benjamin Franklin agreed that each state should have an equal vote in the Senate except in matters concerning money. The convention’s grand committee reported his motion, with some modifications, to the delegates early in July. Madison led the debates against Franklin’s measure, believing it an injustice to the majority of Americans, while some small-state delegates were reluctant even to support proportional representation in the House. On July 16, delegates narrowly adopted the mixed representation plan giving states equal votes in the Senate within a federal system of government.
Once delegates established equal representation in the Senate, they needed to determine how many senators would represent each state. State constitutions offered some guidance. Several states designated one senator per county or district, while in Delaware there were three senators for each of the three counties. Convention delegates did not refer to the state precedents in debate, however. Instead, they seemed to take a common-sense approach in deciding the number of senators.
According to constitutional commentator Joseph Story (1779-1845), few, if any, delegates considered one senator per state sufficient representation. Lone senators might leave their state unrepresented in times of illness or absence, and would have no colleague to consult with on state issues. Additional senators, moreover, would increase the size of the Senate, making it a more knowledgeable body, and better able to counter the influence of the House. On the other hand, a very large Senate would soon lose its distinctive membership and purpose, and actually decrease its ability to check the lower house or to allow senators to take personal responsibility for their actions.
Given these considerations, delegates had a limited choice regarding the number of senators. During the convention, they briefly discussed the advantages of two seats versus three. Gouverneur Morris stated that three senators per state were necessary to form an acceptable quorum, while other delegates thought a third senator would be too costly. On July 23, delegates filled in the blank in the proposal offered by Morris and Rufus King: “That the representation in the second branch consist of _____ members from each State, who shall vote per capita.” Only Pennsylvania voted in favor of three senators. When the question turned to two, Maryland alone voted against the measure, not because of the number, but because Martin disagreed with per capita voting, which gave each senator, rather than each state, one vote.
In its final form, the clause in the Constitution is deceptively simple. “The Senate shall be composed of two senators from each state” appears to be a single provision, the designated number of senators per state. Delegates agreed to this number, however, only after they had considered a larger matter: legislative representation. While representation proved to be the most controversial issue in the convention, delegates determined the number of senators quickly and with little dispute.
The Senate of the United States shall be composed of two Senators from each State, chosen by the Legislature thereof, for six Years. [U.S. Constitution, Article I, section 3, clause 1]
The six-year Senate term represented a compromise between those constitutional framers who wanted a strong, independent Senate and those who feared the possible tyranny of an aristocratic upper house, insulated from popular opinion. Although the six-year term was not utilized in Parliament, Congress under the Articles of Confederation, or the states’ upper houses, these institutions gave the participants at the Constitutional Convention some insight into the impact of term lengths on legislative bodies. While few delegates to the 1787 convention wanted to emulate the House of Lords’ lifelong terms or the Articles of Confederation’s annual appointment of legislators subject to recall, the framers’ reaction against these extremes helped formulate their arguments for and against long terms in the Senate.
Convention delegates seeking information about more moderate term lengths turned to the state constitutions. Although the majority of states set one-year terms for both legislative bodies, five state constitutions established longer terms for upper house members. South Carolina’s senators received two-year terms. In Delaware, the senate had three-year terms with one-third of the senate’s nine members up for reelection each year. New York and Virginia implemented a similar class system but with four-year terms instead of three. Only Maryland’s aristocratic senate featured five-year terms, making this legislative body the focus of the convention’s Senate term debates.
Framers either praised Maryland’s long terms for checking lower house democracy or feared them for the same reason, while some members of the convention believed even five-year terms were too short to counteract the dangerous notions expected to emerge from the House of Representatives. In June, James Madison, Edmund Randolph, and other convention delegates cited Maryland’s experiences when they argued for long Senate terms. According to Madison, Maryland’s senate had never “created just suspicions of danger.” Far from being the more powerful branch, the senate had actually yielded too much, at times, to Maryland’s House of Delegates. Unless the Senate obtained sufficient stability, Madison expected a similar situation under the new Constitution. He suggested terms of seven years or more to counter the influence of the democratic House of Representatives. Randolph believed that the primary object of an upper house was to control the more numerous lower house. He noted that Maryland’s senate had followed this principle but had been “scarcely able to stem the popular torrent.” Seven-year terms, then, had a greater chance of checking the House than terms of five years or fewer.
On June 13, the convention’s Committee of the Whole reported an amended version of Madison’s Virginia Plan. The revised document now designated seven-year Senate terms, a provisional length immediately criticized by some of the framers. For Alexander Hamilton, only lifelong terms could check the “amazing violence & turbulence of the democratic spirit.” Other delegates preferred four-year terms, while Madison began advocating a new nine-year, three-class plan. On June 26, the majority of delegations voted against the nine-year plan, then adopted the six-year Senate term by a vote of 7 to 4.
During the debate on the Constitution’s ratification, Madison became a strong supporter of six-year terms. In the Federalist papers, he argued that Maryland’s experiment with five-year terms proved that slightly longer terms posed no danger to bicameral legislatures. In fact, he expected the six-year terms to have a stabilizing effect on the new national government. Long terms would control turnover in the legislature, allow senators to take responsibility for measures over time, and make senators largely independent of public opinion.
Immediately after they shall be assembled in Consequence of the first Election, they shall be divided as equally as may be into three Classes. [Article I, section 3, clause 2]
Every two years, one-third of the Senate’s members must be reelected or vacate their seats at the end of their terms. Constitutional framers based this three-class system on precedents established by state governments. Delaware’s senate and Pennsylvania’s unicameral council were divided into three classes on a one-year election rotation, while upper houses in Virginia and New York had four classes and yearly elections.
At the 1787 Constitutional Convention, delegates linked the Senate class measure to the debate on term lengths. On June 25, Nathaniel Gorham suggested a four-year Senate term with one-fourth of the senators elected each year. Edmund Randolph supported staggered rotation in the Senate, but wanted a seven-year term “to go out in fixt proportion.” Hugh Williamson countered that six-year terms were more easily divisible into equal election cycles than seven-year terms. The following day, Gorham called for a six-year term, “one third of the members to go out every second year.” Delegates considered a nine-year term, then passed the six-year, three-class Senate clause by a vote of 7 to 4.
Constitutional commentator Joseph Story (1779-1845) explained the reasoning behind the Senate’s class system. Framers hoped biennial elections would bring stability to the Senate, and in turn, to other branches of the new government. By gradually changing members, class rotations would prevent senators from permanently combining for “sinister purposes,” protect the Senate from a rapid turnover in ideas, and encourage senators to deliberate measures over time. Most important, as the federal government’s only continuing body, the Senate could provide leadership after major elections and during other periods of national uncertainty.
At the start of the first session of Congress in 1789, the senators were divided into the three classes by lot with same-state senators assigned to separate groups. The first class’ term expired in two years, the second in four years, and the third in six years. Subsequent elections to all classes were for the full six-year Senate term.
No Person shall be a Senator who shall not have attained to the Age of thirty Years, and been nine Years a Citizen of the United States, and who shall not, when elected, be an Inhabitant of that State for which he shall be chosen. [U.S. Constitution, Article I, section 3, clause 3]
Delegates to the 1787 Constitutional Convention supported establishing membership limitations for House and Senate members. Influenced by British and state precedents, they set age, citizenship, and inhabitancy qualifications for senators, but voted against proposed religion and property requirements.
Age: The constitutional framers debated the minimum age for representatives before they considered the same qualification for senators. Although Delegate James Wilson stated that “there was no more reason for incapacitating youth than age, where the requisite qualifications were found,” other delegates were in favor of age restrictions. They were familiar with England’s law requiring members of Parliament to be twenty-one or older, and they lived in states which either barred legislators under the ages of twenty-one or twenty-five from the upper chambers.
As introduced in May 1787, James Madison’s Virginia Plan left Senate age restrictions to the delegates to decide, only stating that members of the second branch must “be of the age of ____ at least.” Without debating the subject further, delegates voted in favor of filling the blank with thirty, and passed the clause unanimously on June 25, three days after designating twenty-five as the minimum age for representatives. In The Federalist, No. 62, Madison justified the higher age requirement for senators. By its deliberative nature, the “senatorial trust,” called for a “greater extent of information and stability of character,” than would be needed in the more democratic House of Representatives.
Citizenship: Under English law, no person “born out of the kingdoms of England,Scotland, or Ireland” could be a member of either house of Parliament. While some delegates may have admired the “strictness” of this policy, no framers advocated a blanket ban on foreign-born legislators. Instead, they debated the length of time members of Congress should be citizens before taking office. The states’ residency qualifications offered moderate guidelines in this regard. New Hampshire’s state senators needed to be residents for at least seven years prior to election. In other states, upper house members fulfilled a five, three, or one-year requirement, while state representatives completed a residency period of one to three years.
The Virginia Plan made no mention of citizenship when Edmund Randolph introduced it to the convention in May. Two months later, the Committee of Detail reported a draft of the Constitution. Article V, section 3 included a four-year citizenship requirement for senators. On August 9, Gouverneur Morris moved to replace the four-year clause with a fourteen-year minimum. Later that day, delegates voted against citizenship requirements of fourteen, thirteen, and ten years before passing the nine-year provision, making the Senate requirement two years longer than that for the House of Representatives.
At the convention, delegates viewed the nine-year citizenship qualification as a compromise “between a total exclusion of adopted citizens” and an “indiscriminate and hasty admission of them.” While they were concerned that the Senate, especially, might be subject to foreign influence, they did not wish to close the institution to naturalized citizens of merit. Two foreign-born framers expressed these opposing considerations. According to Pierce Butler, recent arrivals were dangerously attached to their countries of origin, a particular concern for senators whose role would include review of foreign treaties. From his own experience, he believed that naturalized citizens would need sufficient time to learn and appreciate American laws and customs before they could serve in government. For James Wilson, however, lengthy citizenship requirements “discouraged and mortified” everyone they excluded. He agreed with Benjamin Franklin that a strict policy would hinder positive immigration and offend those Europeans who had supported the Revolutionary War. On August 13, Wilson moved to reduce the Senate qualification by two years. Delegates rejected his motion, and confirmed the nine-year requirement by an 8 to 3 vote.
Inhabitancy: Although England repealed Parliament’s residency law in 1774, no delegates spoke against a residency requirement for members of Congress. The qualification first came under consideration on August 6 when the Committee of Detail reported its draft of the Constitution. Article 5, section 3 stated, “Every member of the Senate shall be . . . at the time of his election, a resident of the state from which he shall be chosen.”
On August 8, Roger Sherman moved to strike the word “resident” from the House version of the clause, and insert in its place “inhabitant,” a term he considered to be “less liable to misconstruction.” Madison seconded the motion, noting that “resident” might exclude people occasionally absent on public or private business. Delegates agreed to the term, “inhabitant,” and voted against adding a time period to the requirement. The following day, they amended the Senate qualification to include the word, “inhabitant,” prior to passing the clause by unanimous consent.
The Vice President of the United States shall be President of the Senate, but shall have no Vote unless they be equally divided. [U.S. Constitution, Article I, section 3, clause 4]
On September 4, 1787, the term, “vice president,” appeared for the first time at the Constitutional Convention. The Committee of Unfinished Portions reported a method of electing the president and vice president, and recommended that the vice president be the “ex-officio President of the Senate.” Although delegates passed the latter measure three days later, few understood the extent of the vice president’s duties, not having a similar position in their own states. New York’s constitution, however, contained the clear precedent: “the lieutenant-governor shall, by virtue of his office, be president of the senate, and, upon an equal division, have a casting voice in their decisions, but not vote on any other occasion.” New York, then, is aptly credited for the establishment of the vice president’s office in both its executive and senatorial functions.
Convention delegates first considered the selection of the Senate’s presiding officer after the Committee of Detail presented a draft of the Constitution on August 6. Article III, section 4 stated, “The Senate shall choose its own President,” while another article designated the president of the Senate as the executive’s immediate successor. Both provisions passed without dissent. Later in the month, however, increasing support for a separation of powers motivated the Committee of Unfinished Portions to create the electoral college system, rather than have the national legislature elect the president. The committee also suggested that a vice president succeed the executive in the event of a vacancy in that position, but would otherwise serve as the president of the Senate, casting votes only to break a tie. On September 7, Elbridge Gerry and George Mason spoke against the proposed measure, believing that it conflicted with the goal to keep the executive and legislative departments separate and distinct. Roger Sherman defended the clause. “If the vice-President were not to be President of the Senate, he would be without employment, and some member by being made President must be deprived of his vote.” Evidently, most delegates agreed with Sherman’s reasoning, and they passed the measure by an 8 to 2 vote.
In The Federalist, No. 68, New York delegate Alexander Hamilton explained the necessity of the vice president’s Senate position. To secure definitive resolutions, the Senate’s president must be able to cast tie-breaking votes, yet be denied a vote at all other times. Therefore, the Senate’s presiding officer must not be a member of the Senate, nor should a senator be next in line for the presidency, since the president’s successor should be chosen in the same manner as the president.
The Senate shall chuse their other Officers, and also a President pro tempore, in the Absence of the Vice President, or when he shall exercise the Office of President of the United States. [Article 1, section 3, clause 5]
Before Constitutional framers designated the vice president as the president of the Senate on September 7, 1787, they granted senators the right to choose other Senate officers, including those from outside the elected body. The clause was not debated at the convention, but simply assumed in the Committee of Detail’s report to the Constitutional Convention on August 6. According to James Madison, numerous precedents made the measure “so obvious that it [was] wholly unnecessary to vindicate it.” Members of Parliament had been selecting clerks of the house and sergeants at arms for hundreds of years, and state legislatures appointed various administrative officers. The Senate modeled its own offices of the secretary, the sergeant at arms, and the doorkeeper after positions established in the Continental Congress.
Unlike the secretary and the sergeant at arms, the president pro tempore is an elected member of the Senate, chosen by the Senate to preside in the absence of the vice president. The position resulted from the framers’ decision to make the vice president the president of the Senate, but it originated in the state legislatures. While other states provided stand-ins for presiding officers, New York specified the arrangement in its constitution: whenever the lieutenant-governor “shall be unable to attend as president of the senate, the senators shall have power to elect one of their own members to the office of president of the senate, which he shall exercise [until the lieutenant governor resumes his position in the chamber].”
At the convention, delegates first considered having an elected senator be the president of the Senate, as shown in the Committee of Detail’s report on August 6. However, in September, Roger Sherman observed that, except in the case of a tie, the presiding officer would be deprived of his vote. Rather than disadvantage one state by elevating its senator to a non-voting position, framers made the vice president the president of the Senate. As of September 10, they accepted the president pro tempore as the presiding officer’s substitute to preside over the Senate, or select another senator to preside, during the absence of the vice president.
The Senate shall have the sole Power to try all Impeachments . . . And no Person shall be convicted without the Concurrence of two thirds of the Members present. [Article I, section 3, clause 6]
Early in the 1787 Constitutional Convention, most delegates agreed that the inclusion of an impeachment provision would help to hold national officers accountable for their actions. The Senate’s role in impeachment trials, however, developed after months of consideration behind the closed doors of committee rooms. Based on those of the British Parliament and the state constitutions, the Senate impeachment provision gave senators the responsibility for trying impeached officials, including the president of the United States.
Throughout the summer of 1787, committee members reported impeachment plans to the full convention. The preliminary resolutions were considered by the Committee of the Whole and returned to selected delegates for further revision. First submitted on May 29, James Madison’s Virginia Plan proposed a supreme tribunal to hear and determine cases including, among other concerns, the “impeachments of any National officers.” On June 13, the Committee of the Whole amended the plan’s proposition. Section nine of the committee’s report stated that the president could be “removable on impeachment of malpractices or neglect of duty.” The revised measure did not specify the procedures for trying the national executive.
In June and July, the framers debated the merits of involving Congress in the impeachment process. Roger Sherman “contended that the National Legislature should have the power to remove the Executive at pleasure.” George Mason objected to Sherman’s plan, claiming that the president would become the “creature of the Legislature.” John Dickinson countered with an unsuccessful motion to make the executive “removable by National Legislature at request of majority of State Legislatures.”
On August 6, the Committee of Detail reported that the House of Representatives “shall have the sole power of impeachment” and the executive “shall be removed from his office by conviction in the supreme Court, of treason, bribery, or corruption.”Two weeks later, the committee added that “the judges of the supreme court be triable by the senate, on impeachment by the house of representatives.”
The constitutional plan then went for review to the Committee of Eleven, consisting of one member from every state represented at the convention. Presented to the full convention on September 4, the Committee of Eleven’s report stated, “The Senate of the U.S. shall have power to try all impeachments [by the House of Representatives]; but no person shall be convicted without the concurrence of two thirds of the members present.” The framers debated the clause on September 8. Arguing that the executive would become dependent on the legislature, Madison opposed Senate impeachment trials. He moved to strike out the words “by the Senate” after the word “conviction,” but the resolution was rejected. Later that day, the delegates agreed to Gouverneur Morris’ addition, “and every member shall be on oath” before they passed the final measure by an 8 to 2 vote.
In The Federalist, No. 65, Alexander Hamilton explained the Committee of Eleven’s rationale for placing impeachment trials in the Senate: no other institution would be sufficiently dignified or independent to handle the proceedings. Furthermore, the British Parliament and the state constitutions provided similar models for legislative impeachments. In England, impeachments were instituted by the House of Commons and tried by the House of Lords. Penalties for conviction ranged from fines to jail, banishment, or death. Following the Declaration of Independence, the states based their constitutions’ impeachment clauses on their colonial charters, which limited both the punishment and the conditions for impeachment.
Madison, in The Federalist, No. 47, referred to the impeachment provisions in the New York, New Jersey, and Massachusetts state constitutions. In New York, members of the legislature and the judiciary served on a court of impeachment, while in New Jersey, select officers could be dismissed by the upper house on impeachment by the lower house. The clearest antecedent to the U.S. impeachment clause, however, is found in Massachusetts’ 1780 constitution. Section 2, Article VIII states, “The senate shall be a court, with full authority to hear and determine all impeachments made by the house of representatives, against any officer or officers of the commonwealth, for misconduct and maladministration of their offices.”
He shall have Powers, by and with the Advice and Consent of the Senate, to make Treaties, provided two thirds of the Senators present concur . . . . [U.S. Constitution, Article II, section 2, clause 2]
As they debated the controversial treaty-making clause, the delegates to the 1787 Constitutional Convention considered, but did not follow in whole, those precedents with which they were most familiar. In Great Britain, treaties were made by the king and, in certain cases, had to be approved by a majority vote in Parliament. The Continental Congress, which had no executive branch, dispatched agents to negotiate treaties. The treaties only went into effect after two-thirds (nine out of thirteen) of the states approved the documents. This inefficient process was further complicated by the states’ ability to enter into their own treaties. While the delegates agreed that the states could not continue to make treaties with foreign powers, they disagreed over the manner in which the United States should negotiate, draft, and ratify international pacts.
On August 6, the convention’s Committee of Detail reported a preliminary Constitution to the full convention. Article IX, section 1 stated, “The Senate of the United States shall have power to make treaties, and to appoint Ambassadors, and Judges of the Supreme Court.” Throughout August and into the month of September, the delegates debated treaty-making as a separate issue from the rest of the clause. Several delegates opposed granting the Senate sole control over treaty-making. While some wanted the executive to have that responsibility, others advocated involving the entire legislature in the process. Small-state delegates, however, preferred the Committee of Detail’s treaty clause because it gave each state an equal say in the adoption or rejection of treaties.
On September 4, the Committee of Eleven reported a treaty clause that appeased many of the delegates: “The President by and with the advice and Consent of the Senate, shall have power to make Treaties.” After further debate, the delegates unanimously approved the clause on September 7. The clause, however, was taken up again, this time to add to it the words, “But no treaty shall be made without the consent of two thirds of the members present.” Shortly thereafter, the convention adopted James Madison’s addition, “except in treaties of peace,” which would be ratified by a simple majority vote. The next day, the delegates struck out the peace treaty exception and considered dropping the Senate super majority as well. After two delegates cited the Continental Congress’ “two-thirds of the states” example, however, they voted to keep the two-thirds requirement.
Although adopted by the convention, the treaty clause continued to stir debate in the period before the Constitution’s ratification. As one of the clause’s strongest proponents, Alexander Hamilton defended the provision in The Federalist, No. 75. Remarkably, given the delegates’ extreme dissension over treaty-making, he wrote, the clause “is one of the best digested and unexceptionable parts of the plan.”
* * * * *
[The president] shall nominate, and by and with the Advice and Consent of the Senate, shall appoint Ambassadors, other public Ministers and Consuls, Judges of the supreme Court, and all other Officers of the United States . . . . [U.S. Constitution, Article II, section 2, clause 2]
Debated over the course of several weeks, the Constitution’s nomination clause split the delegates into two factions: those who wanted the executive to have the sole power of appointment, and those who wanted the national legislature, and more specifically, the Senate, to have that responsibility. The latter faction followed precedents established by the Articles of Confederation and most of the state constitutions. These documents granted the Continental Congress and the state legislatures the power to make appointments. The Massachusetts constitution provided an alternative model, however. For over one hundred years,Massachusetts had divided the appointment responsibilities between its governor, who made the nominations, and its legislative council, which confirmed the appointments.
Rather than adopt the Massachusetts model immediately, the convention delegates initially granted the president the power to appoint the officers of the executive branch and, given that judges’ life-long terms would extend past the authority of any one president, the Senate would appoint the members of the judiciary. Framers in favor of a strong executive, however, argued that Senate appointments would lead to government by a “cabal” swayed by the interests of constituents. Other delegates, fearful of monarchies, wanted to remove the president entirely from the appointment process. On September 4, the Committee of Eleven reported an amended appointment clause. Unanimously adopted on September 7, the clause, based on the Massachusetts model, provided that the president shall nominate and, with the advice and consent of the Senate, appoint the officers of the United States.
Anderson, Thorton. Creating the Constitution: The Convention of 1787 and the First Congress. University Park, Penn.: The Pennsylvania State University Press, 1993.
Hamilton, Alexander, James Madison, and John Jay. The Federalist Papers. Edited by Clinton Rossiter. New York: Mentor, 1999.
Haynes, George. The Senate of the United States. Boston: Houghton Mifflin Company, 1938.
Jillson, Calvin C. Constitution Making: Conflict and Consensus in the Federal Convention of 1787. New York: Agathon Press, 1988.
Kurland, Philip B. and Ralph Lerner. The Founders’ Constitution. 4 volumes. Chicago: University of Chicago Press, 1987.
Madison, James. Notes of Debates in the Federal Convention of 1787. Athens, OH: Ohio University, 1966.
Story, Joseph. Commentaries on the Constitution of the United States. 2 volumes, fifth edition. Edited by Melville Bigelow. Boston: Little, Brown, and Company, 1891.
Swindler, William F., ed. Sources and Documents of United States Constitutions. 10 volumes. Dobbs Ferry, NY: Oceana Publications, Inc., 1973-1979.
Van Doren, Carl. The Great Rehearsal: The Story of the Making and Ratifying of the Constitution of the United States. New York: The Viking Press, 1948.