[LETTERHEAD OF WILLKIE FARR & GALLAGHER LLP]



VIA EDGAR
- ---------


May 25, 2006


Securities and Exchange Commission
100 F Street, N.E.
Judiciary Plaza
Washington, D.C.  20549
Attn:  Jim B. Rosenberg, Senior Assistant Chief Accountant
       Division of Corporation Finance


Re:  W. R. Berkley Corporation
     Form 10-K for the Fiscal Year Ended December 31, 2005
     Filed March 13, 2006
     File No. 1-15202

Dear Mr. Rosenberg:

     On behalf of W. R. Berkley Corporation ("Berkley" or the "Company"), set
forth below are Berkley's responses to the comment of the staff (the "Staff") of
the Securities and Exchange Commission (the "Commission") provided orally on
April 27, 2006. For your convenience, we have set forth below a summary of the
Staff's comment in italics, followed by Berkley's response thereto.

Form 10-K for the Fiscal Year Ended December 31, 2005 (the "2005 Form 10-K")
- ----------------------------------------------------------------------------

Management's Discussion and Analysis of Financial Condition and Results of
- --------------------------------------------------------------------------
Operations
- ----------

Reserves for Losses and Loss Expenses, page 3
- ---------------------------------------------

1.   We note your response to comment 2 in our letter dated March 23, 2006, in
     which you state that it would not be useful to investors to present your
     reserve sensitivity analysis on a disaggregated basis. We believe that a
     disaggregated analysis would be more meaningful to investors because the
     key assumptions are different for short-tail and long-tail businesses.
     Please provide us in disclosure-type format a revised sensitivity analysis
     by tail or line of business and link that analysis to the discussion of
     your key assumptions for each tail or line of business. This discussion
     should focus on the reasonably likely changes in the key assumptions that
     caused management to arrive at the volatility presented.




Mr. Jim B. Rosenberg
Securities and Exchange Commission
May 25, 2006
Page 2


     In response to this comment, Berkley proposes to add the below disclosure,
     commencing with its Quarterly Report on Form 10-Q for the Quarterly Period
     Ended June 30, 2006. Rather than adding separate sensitivity tables,
     distinguishing Berkley's short and long-tail businesses, Berkley proposes
     to enhance its description of the factors inherent in computing loss
     reserves, in particular the different impact of loss reporting lag on its
     reserves and the key assumptions involved in such analysis. We believe that
     a second similar table would merely clutter, rather than enhance, the
     presentation of this information.

     Instead, as set forth in the proposed language below, Berkley believes that
     it would be more useful to provide investors with historical data
     indicating the variability of Berkley's reserves from its initial
     estimates, broken out by lines with long and short reporting lags. This
     additional information would provide investors with the specific
     information to evaluate the sensitivity of Berkley's loss reserves
     estimates.

         One factor affecting estimates of loss frequency and severity is the
         loss reporting lag, which is the period of time between the occurrence
         of a loss and the date the loss is reported to the Company. The length
         of the loss reporting lag affects loss frequency (loss frequencies are
         more predictable for lines with short reporting lags) as well as the
         amount of reserves needed for incurred but unreported losses ("IBNR")
         (less IBNR is required for lines with short reporting lags). As a
         result, loss reserves for lines with short reporting lags are likely
         to have less variation from initial loss estimates. For lines with
         short reporting lags, which include commercial automobile, primary
         workers' compensation, commercial multi-peril business, other
         liability (claims-made) and property business, the key assumption is
         the loss emergence pattern used to project ultimate loss estimates
         from known losses paid or reported to date. For lines of business with
         long reporting lags, which include other liability (occurrence),
         products liability, excess workers' compensation and liability
         reinsurance, the key assumption is the expected loss ratio since there
         is little paid or incurred loss data to consider.

         Historically, the Company has experienced less variation from its
         initial loss estimates for lines of businesses with short reporting
         lags than for lines of business with long reporting lags. For example,
         as of December 31, 2005, initial loss estimates for accident years
         1996 through 2004 were increased by an average of 6% for lines with
         short reporting lags and by an average of 37% for lines with long
         reporting lags. For the latest accident year ended December 31,



Mr. Jim B. Rosenberg
Securities and Exchange Commission
May 25, 2006
Page 3

         2005, initial loss estimates were $1.5 billion for lines with short
         reporting lags and $1.0 billion for lines with long reporting lags.

    On behalf of Berkley, we hereby acknowledge that:

    o    Berkley is responsible for the adequacy and accuracy of the disclosure
         in its 2005 Form 10-K;

    o    Staff comments or changes to disclosure in response to Staff comments
         do not foreclose the Commission from taking any action with respect to
         Berkley's 2005 Form 10-K; and

    o    Berkley may not assert Staff comments as a defense in any proceeding
         initiated by the Commission or any person under the Federal securities
         laws of the United States.

     Should any member of the Staff have any questions or comments concerning
this letter, please do not hesitate to call me at (212) 728-8592.

Very truly yours,

/s/ Jeffrey S. Hochman

Jeffrey S. Hochman

Enclosures

cc:  Lisa K. Vanjoske
     Dana Hartz
     Eugene G. Ballard
     Ira S. Lederman, Esq.